How Healthcare Chatbot Service Failures Affect Patient Trust?

How Healthcare Chatbot Service Failures Affect Patient Trust? | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article How Healthcare Chatbot Service Failures Affect Patient Trust? Shanping Hu, Dajun Yang, Gui Gui, Yunyun Wei, Jun Cao, Daibo Xiao This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6930952/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study aimed to investigate the impact of healthcare chatbot service failure on patient trust and examined the role of two factors, patient mindset and the linguistic expression type of the healthcare chatbot service. By developing a theoretical framework and conducting the evidence-based experiment, this study argued that healthcare chatbot service failure reduces patient trust in the robot, by affecting patient mindset, as well as changing patients' views and perceptions of the healthcare chatbot service. This study first presented the impact of healthcare chatbot service failure on patients' trust shaping, and explored the interaction effect on patients' trust, from the perspectives of patients' mindset and the linguistic expression type of the healthcare chatbot service. The contribution of this study is to fill the research gap on the impact of the type of healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, and to explore the interaction effect between patient mindset and the linguistic expression type of the healthcare chatbot service and its operational mechanisms. These findings are important for the understanding and application of healthcare chatbot service and patient trust, providing valuable guidance for the development of healthcare chatbot service models, the maintenance of doctor-patient relationships, and service capabilities of hospital. Biological sciences/Psychology Biological sciences/Psychology/Human behaviour healthcare chatbot patient trust service failure linguistic expression type 1 INTRODUCTION The healthcare chatbot is a multifunctional intelligent digital healthcare device (Kakhi et al., 2022 ). In addition to providing functions such as professional surgery(Han et al., 2022 ), postoperative rehabilitation guidance, and other auxiliary services(Guo et al., 2021 ; Han et al., 2022 ), healthcare chatbots also play an important role in the patient care process (Guo & Li, 2022 ). There has been a great deal of resources invested in researching how to enhance the service of healthcare chatbots to help healthcare personnel reduce working pressure(Alhaffar & Janos, 2021 ) and meet the needs of patients in accessing healthcare. In recent years, the pressure on healthcare organizations and the workload of medical workers have increased dramatically, making it particularly important to actively utilize the service of healthcare chatbots (Guo et al., 2021 ). It has been told that healthcare chatbot service failure can have a huge impact on patients' psychology and attitudes(Choi et al., 2021 ). However, it is also reasonable to doubt whether the different types of healthcare chatbot service failure (process failure vs. outcome failure) actually affect patient trust in the healthcare chatbot. Nevertheless, academics in the field of robotic healing care have presented limited evidence to respond to this question. Robotic service failure is a situation where the robot has problems performing a task or providing a service or is unable to complete the task. An extensive literature exists on the topic of healthcare chatbot services, and it has become a trend in research to promote the widespread healthcare chatbots (Gibelli et al., 2021 ; Lee et al., 2020 ). Although previous studies have highlighted that patient trust could influenced by the therapeutic effectiveness and nursing of healthcare chatbots (Liao et al., 2023 ; Liu et al., 2022 ). For example, in a study by Torrent-Sellens et al. ( 2021 ), it was found that patient trust in healthcare chatbot therapy overrides rational decision-making. However, research on how healthcare chatbot service failure affects patient trust is somewhat flawed and insufficient. A recent study suggested that the healthcare chatbot can improve the way of patients trust in healthcare (Kerasidou, 2020 ). Although previous studies have focused on the impact of healthcare chatbot service on patient trust (Asan et al., 2020 ; Holthöwer & van Doorn, 2023 ), not enough attention has been paid to whether healthcare chatbot service failure affects patient trust (Leo & Huh, 2020 ), so its value in terms of theory and nursing care has not been adequately recognized. This neglect is particularly unfortunate because healthcare chatbot service failure may trigger patients' skepticism towards healthcare organizations and question the overall competence of the chatbot, thus causing patients' disappointment, frustration and anxiety, which could ultimately affect therapeutic outcomes and the recovery process (Khaksar et al., 2024 ; Shanks et al., 2024 ). More importantly, in the field of healthcare chatbot service failure, the mechanism and effect of the type of healthcare chatbot service failure on patient trust have not been demonstrated clearly. Therefore, we need to answer the following questions: How does healthcare chatbot service failure affect patient trust? What are the internal mechanisms by which healthcare chatbot service failure affects patient trust? To fill the gap of previous research, this study developed a theoretical framework to investigate how to restore patient trust in healthcare chatbot service failure scenarios. Guided by the semantic framing, we argued that healthcare chatbot service failure reduces patient trust by affecting their mindset and altering their views and perceptions of the healthcare chatbot service. Despite previous research on the ability of healthcare chatbot service to influence patients' perceptions (Koutentakis et al., 2020 ), insufficient attention has been paid to the impact of healthcare chatbot service failure on patient trust. Based on this, we first proposed the impact of healthcare chatbot service failure on patient trust; second, we verified the interaction effect between patient mindset and healthcare chatbot service failure on patient trust; Last we explored the interaction effect between the linguistic expression type of the healthcare chatbot and healthcare chatbot service failure on patient trust. This study makes the following important contributions to service robots and patient trust. Although it has been common for studies to address healthcare chatbot service (Choi et al., 2021 ), few studies have focused on the types of healthcare chatbot service failure and their impact on patient trust. Therefore, this study is an early attempt to link the healthcare chatbot service failure with the patient trust, which further clarified the internal mechanisms and boundary conditions between the two, enriching the existing literature on healthcare chatbot services and patient trust. Second, this study explored the interaction effect between patient mindset (Cowger et al., 2020 ) and healthcare chatbot service failure. Patients' growth mindset and fixed mindset towards the healthcare chatbot service may have differential effects on patient trust, and this could also better explain the impact of changed patient mindset on patient trust with healthcare chatbot service failure. Last, this study examined the interaction effect between the linguistic expression type of the healthcare chatbot service and healthcare chatbot service failure. Robot language is mostly set by programming algorithms, but the linguistic expression type significantly affects patient trust (Foster, 2019 ), and an in-depth exploration could better illustrate the internal mechanisms between healthcare chatbot service failure and patient trust. In summary, the theoretical contribution of this study is to provide a systematic explanation and understanding of how healthcare chatbot service failure combines with patient mindset and linguistic expression type, which in turn affects patient trust. In fact, this study also provides valuable guidance for the development of healthcare chatbot service models, the maintenance of doctor-patient relationships, and the service capabilities of hospitals. 2 THEORETICAL FRAMEWORK AND HYPOTHESIS DERIVATION 2.1 AI AND SERVICE ROBOTS With the continuous development of science and technology, robots are increasingly used in various fields. Especially in the service industry, robots have become an important tool for providing efficient and convenient services (Brooks, 1991 ; Luo et al., 2024 ). At the same time, the drawbacks of robot services have gradually emerged and attracted extensive attention from academics (Moriuchi & Murdy, 2024 ; Ryoo et al., 2024 ). Based on the related literature, first we categorized the current literature on the connotation of AI services, and then provided an overview of the existing research in terms of both positive and negative factors that affect consumers' adoption of AI in the service industry. 2.1.1 AI SERVICE A large amount of literature has discussed the connotation of AI services. The definition of AI services is rooted in AI technology. According to Liu et al. ( 2018 ), AI refers to intelligent programs, algorithms, systems, and machines, which can present different characteristics of human intelligence (Holzinger et al., 2019 ). Huang et al. ( 2023 ) specified 4 types of intelligence of AI, including mechanical, analytical, intuitive and empathetic. Based on this, Doupe et al. ( 2019 ) defined robot services as system-based autonomy and adaptability that can interact, communicate and deliver services to organized customers. Researchers also classified service robots according to the dimensions of functionality, socio-emotion, and relationship. Recently, Bock et al. ( 2020 ) tend to identify AI services as a technological configuration that deliver service value through perception, learning, decision-making, and action realization with flexible adaptation. Integrating the existing literature, this study argued that AI services can be defined as follows. It is a process by which algorithm- and technology-driven intelligent devices (which may have anthropomorphic appearances, languages and personalities) provide service value to customers with communication and interaction. With the continuous improvement and maturation of AI technology, future AI services will be presented with trend of intelligence, personalization and emotionalization. 2.1.2 SERVICE ROBOT Service robots is a relatively new field whose definition and scope are evolving over time(Chi et al., 2023 ). Taking into account different definitions, service robots are described as robots that are able to perform useful tasks. These performed tasks can either serve humans or devices (Becker et al., 2023 ). The development of service robots has reduced the workload of healthcare workers and helped them carry out complex tasks. The earliest definition of a service robot was proposed by Schraft in 1993. They defined it as “A service robot is a freely programmable kinematic device that performs services semi-or fully automatically. Services are tasks that do not contribute to the industrial manufacturing of goods but are the execution of useful work for humans and equipment” (Schraft et al., 1993 ). As service robots have evolved, researchers have come up with additional definitions. The International Organization for Standardization defines service robots as “a robot that performs useful tasks for humans or equipment, excluding industrial automation applications”. The International Federation of Robotics, on the other hand, emphasizes the autonomy of robots, defining service robots as “a service robot is a robot which operates semi- or fully autonomously to perform services useful to the well-being of humans and equipment, excluding manufacturing operations”(Holland et al., 2021 ). However, as the term "Service Robot" continues to evolve, the its definition has become ambiguous, because there is a crossover between the industrial and service sectors (Dupont et al., 2021 ). To sum up, this study argued that a service robot is capable of performing useful tasks that is intended to provide services to humans and devices. The definition and scope of service robots have evolved over time, but their core goal has always been to deliver helpful services to humans and devices. 2.1.3 FACTORS AFFECTING THE APPLICATION OF SERVICE ROBOT Existing literature focused on how AI services affect customer psychology and thus act on customer behavior from the perspective of human-machine interaction. Part of those literatures explored the positive incentives that motivate consumers to adopt AI from different aspects, including consumer trust in anthropomorphic robots, and their perception of anthropomorphic robots' enthusiasm. Specifically, Chuah et al. ( 2021 ) demonstrated experimentally that having more anthropomorphic features in consumer robots would increase customers' perception of enthusiasm. Trust in intelligent algorithms was higher when humans performed those less threatened tasks. An experimental study by Barone et al. ( 2024 ) demonstrated that anthropomorphism has a function in guiding consumers to make positive evaluations of robots when they have failed to provide service. Because customers' anthropomorphic perceptions of service robots facilitated their intention to use the service robots. It was found that anthropomorphic robots can repair customers' enthusiasm from service remediation through apologies and explanations, which in turn leads to renewed satisfaction, and consumer trust enhances their willingness to use humanoid service robots (Hwang et al., 2020 ). Although existing studies have found an "algorithm aversion" (Reich et al., 2023 ), some academic researchers have suggested that the positive impact of anthropomorphism on transaction switching has been progressively explored, as consumers' increasing acceptance of the algorithms in their daily lives (Roesler et al., 2023 ). There were also studies discussed the negative factors that affect customers' adoption of AI, such as the lack of compassion and empathy of AI and consumers' lack of control over AI services(Dang & Liu, 2022 ). The discomfort may be caused by overly anthropomorphic robots and the consumers' personality could be neglected by AI(Zebrowski, 2020 ). Meanwhile, the literature has found that compassion and empathy are two human qualities that are lacking in machines. Therefore, customers' distrust, feelings of being ignored and unappreciated, and closed mind that resist change also limit the adoption of AI services (Kuchenbrandt et al., 2013 ). In an earlier study, Mende et al. ( 2019 ) found that consumers of anthropomorphic service robots, even when faced with flawless AI and robotic services, may be affected by speciesism and consequently avoid the use of service robots. In addition, it has also been found that service robots have more challenges than humans in interpreting consumers' unique characteristics and environments in providing healthcare services, which has led to consumers' resistance to healthcare AI services (Huang et al., 2023 ). 2.2 ROBOT SERVICE FAILURE Robot service failure can be defined as any form of actual or perceived misfortune, error, or problem that occurs during the stage of consumer experience (Arikan et al., 2023 ). This failure results in the behavior or service performed by the robotic system deviating from the desired, normal or correct function (Zhang et al., 2023 ). Robot service failures include both perceivable failures and actual failures caused by correct behaviors performed by the robot in accordance with its programming (Hu et al., 2021 ; Liu et al., 2023 ). Robot service failures are inevitable, because various weaknesses in its hardware, structure, and applications can lead to service failures. These weaknesses may include technical limitations, design flaws, and software bugs of the robot. These issues can trigger consumer dissatisfaction and negative word-of-mouth(Leo & Huh, 2020 ). In order to specifically classify robot service failures, existing research has mainly divided it into two dimensions, process failure and outcome failure in interpersonal service contexts(Choi et al., 2021 ). Process failure refers to deficiencies in service delivery, such as untimely or incorrect robot responses. Outcome failure, on the other hand, refers to the failure of the robot to fulfil the basic service content, such as delivering the wrong goods. By dividing robot service failure into process failure and outcome failure, we can better understand the problems in robot service and take appropriate measures to improve service quality. For process failure, we can improve the service efficiency by upgrading robots' response speed and accuracy. For outcome failures, we can enhance the basic service content of the robot to ensure that consumers receive the services they expect(P. X. Wang et al., 2023 ; X. Wang et al., 2023 ). By categorizing different types of failures and proposing specific solutions, we can improve the quality of robot services and user experience. However, robot service failure may have more serious consequences in the healthcare domain as it relates to the health and lives of patients. Therefore, discussion on healthcare chatbot service failure is particularly important. Unfortunately, previous studies have not deeply discussed the antecedents of healthcare chatbot service failure (process failure vs. outcome failure), which has led to the absence of a clear answer in the academic community, on how to regain patient trust in a context of healthcare chatbot service failure. This study inferred that process failures can usually be corrected, whereas outcome failures are probably unchangeable. When patients realize that process failures can be fixed, patients will have more confidence and patience to wait for it to be resolved, rather than feeling despair and disappointment due to outcome failures. In addition, the process failure can be handled in a way that the doctor and robot are open to the patients with transparency and honesty, about the cause of the problem and the solution. It can increase the patient trust in the doctor or robot and avoid disappointment. Based on the above analyses, the following hypothesis was proposed. H1: Healthcare chatbot process failure (vs. outcome failure) has less impact on patient trust. 2.3 INTERACTION EFFECT OF PATIENT MINDSET Patient mindset and robot service failure are two factors that are closely related to patient trust. Patient mindset refers to the patient's mental state and attitude in receiving healthcare services, while robot service failure refers to the problems or mistakes made by the robot in providing healthcare services (Chiu et al., 1997 ). The interaction between these two factors has a significant impact on the emergence and maintenance of patient trust. The interaction between patient mindset and robot service failure on patient trust is a complex process with multiple levels and dimensions (Meng et al., 2008 ). Patient mindset can be broadly categorized into growth mindset and fixed mindset (Chiu et al., 1997 ), while robot service failure includes both process failure and outcome failure. The interaction between these two plays a crucial role in the emergency and maintenance of patient trust. First, we explored the interaction between growth patient mindset and robot service failure. A growth patient mindset implies that patients tend to view the healthcare process positively, believing that through continuous learning and effort, they can overcome the difficulties associated with their illness (Fissell et al., 2021 ). Patients with this mindset are more likely to view process failures of robot services as opportunities to learn and improve rather than as mere failures when faced with them (Grant, 2023 ). They will understand and accept the limitations in its initial stages, while believing these issues will be resolved as the technology continues to advance (Ching et al., 2023 ). As a result, patients with growth mindset are more likely to maintain a high level of trust even if there are process failures in robot services. However, even patients with growth mindset may face trust challenges when there are robot services outcome failures, as this is directly related to their treatment outcome and health status, which is their crucial concern (Joseph et al., 2021 ). It is also difficult for patients who hold a growth mindset to accept failed treatments or deterioration of their condition due to errors in robot services(Chao et al., 2023 ). In this case, patients may become skeptical about the reliability and effectiveness of the robots, thus affecting their trust. Next, we analyzed the interaction between fixed patient mindset and robot service failure. Fixed patient mindset manifests itself as patients holding more fixed views and expectations about their health conditions and medical processes. Such patients may be more prone to negative emotions and distrust when faced with process failures of robot services. They may view process failure as an inherent flaw rather than a normal part of the technological development process(Richardson et al., 2021 ). Thus, for patients with fixed mindset, their trust in robot services may be significantly reduced by process failure. As for outcome failure, the impact may be even more severe for patients with fixed mindset. Such patients usually have high expectations for medical outcomes and may feel extremely disappointed and angry when the robot service fails to achieve the desired treatment outcome (Desveaux & Ivers, 2024 ). They may blame the outcome failure on the deficiencies of the robotic technology and thus lose trust in the robot service altogether. Based on the above analyses, the following hypothesis was proposed. H2: Healthcare chatbot service failure (process failure vs. outcome failure) and patient mindset (growth vs. fixed) have an interaction effect on patient trust. 2.4 SEMANTIC FRAME OF “WANT” VS. “NEED” The semantic frame is a structure used to represent the relationships between words and sentences in language, which describes the semantic relationships between lexical items, including hypernym and hyponym, parts and wholes, attributes and entities, and so on. Semantic frames can help us understand and interpret meaning in our language. In recent years, a large number of academic researchers have focused on the impact of semantic frames on consumer behavior and have demonstrated its prominent importance in marketing and advertising. By choosing the right words or phrases, marketers can convey different emotions and intentions to influence consumers' purchasing decisions and attitudes. For example, when recommending a product or service, using "I recommend it" is more persuasive than "I like it" (Packard & Berger, 2017 ). The former implies a higher level of professionalism and reliability, whereas the latter only expresses personal preferences. Similarly, when asking for help or support, the use of "will you" is more likely to receive a positive response than "can you" (Jia et al., 2021 ). The former implies the other person's willingness and ability, whereas the latter only expresses the other person's ability. These small semantic changes may have far-reaching effects on consumers. Therefore, marketers need to choose and use words carefully to ensure that they convey the emotions and intentions they want. From the findings of psycholinguistic research, it is clear that language can effectively influence people's perceptions of certain people or things. Research has shown that frequent use of the first-person singular has been shown to be neurotic (Pennebaker et al., 2003 ). The research on event-related potentials provides direct neuroscientific evidence that semantics affects individuals' behavior. In recent years, academic researchers have found that semantic symbols can affect individuals' memory, which in turn influences how they act (Seo et al., 2023 ). As a result, semantic associations have gained the attention of marketing scholars. Qian and Yamada ( 2020 ) found that semantic associations have a direct correlation with consumers' attitudes towards products. There are indications that semantic frames can effectively influence individual behavior. Based on findings from the literature on semantic frame, we questioned to what extent a healthcare chatbot dialogue request framed as "need" vs. "want" affects patient trust–"need" vs. "want" triggers different patients' perceptions regarding the professionalism of the healthcare chatbot (Su et al., 2024 ). We further argued that due to these different perceptions, the semantic frames of "need" and "want" have different mechanisms for influencing patient trust. In research on consumption, academics generally supported the idea that "need" refer to the things that consumers must have in order to maintain their basic needs in life. These "need" are the items that people must purchase in order to satisfy their basic survival and living needs, such as food, clothing, shelter, and so on. It can be seen that needs exist to satisfy the basic requirements of life, without which people's lives would be difficult or even impossible to maintain. "Want", on the other hand, refers to consumers' desire or longing for certain items or experiences. These are non-essential items or experiences that do not exist to fulfil basic life needs. The wanted things are usually like luxury items, recreational activities and travelling. They are not essential to life, but can provide additional enjoyment and satisfaction (Su et al., 2024 ). Therefore, the difference between "need" and "want" lies in the fact that "need" is something that must be possessed in order to satisfy the basic necessities of life, whereas "want" is a desire for a non-essential object or experience. "want" is a desire for non-essential objects or experience. A "need" is basic, while a "want" is additional. In consumption decision-making, rational consumers will first satisfy their needs before considering meeting their wants (Baker, 2018 ). This stratification suggests that the "need" and "want" sent to the patient, when the healthcare chatbot service failure occurs, may affect patients' willingness for further interaction. If the healthcare chatbot is able to satisfy the patients' basic needs ("need"), the patient may be more willing to continue interacting with the robot because they perceive their ability to provide the necessary help and support. On the contrary, if the healthcare chatbot only satisfies the patients' additional needs ("wants"), patients may perceive their services as not necessary, thus diminishing patients' willingness for further interaction. On the basis of these considerations, we proposed the central premise of this study: the use of "need" (or "want") in patient interactions with a healthcare chatbot will affect the extent to which the patient trust in the healthcare chatbot is perceived differently. As assumed in the previous discussion, "need" and "want" requests do not imply the same constraints. The "want" semantic frame implies greater adequacy and higher capacity for action without constraints. Therefore, we argued that the use of "need" vs. "want" in healthcare chatbot requests will affect patients' different inferences about the trustworthiness of the healthcare chatbot. (To our knowledge) this inference has never been directly confirmed. However, there is some indicative evidence for healthcare chatbot-patient interactions. For example, Su et al. ( 2024 ) found that "need" vs. "want" affects differences in customer perceptions, which can lead to different individual willingness for donation. The current research tested the difference in patient trust in a healthcare chatbot service failure scenario, after sending a sentence containing "need" vs. "want" to a patient. We argued that it affects patient trust in the healthcare chatbot because the semantic frames of "need" and "want" imply different semantic signals. It is important to reiterate that, consistent with past research, "need" and "want" in the current study are not the result of a binary classification, but rather a continuum. Thus, "need" does not imply an absolute service relationship between the healthcare chatbot and the patient, whereas "want" implies a more conciliatory tone for the healthcare chatbot. Argument in this study is that "want" positively affects patient trust more than "need". Based on the above analyses, the following hypothesis was proposed. H3: Healthcare chatbot service failure (process failure vs. outcome failure) and linguistic expression type (need vs. want) have an interaction effect on patient trust. 3 OVERVIEWS OF THE STUDY To test the above three hypotheses, this study verified the impact of healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, along with its internal mechanisms and boundary conditions, by conducting three related experiments. Experiment 1 verified the main effect of healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, verifying H1; Experiment 2 attempted to investigate the interaction effect between patient mindset (growth vs. fixed) and healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, verifying H2; and Experiment 3 analyzed the interaction effect between linguistic expression types (need vs. want) and healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, verifying H3. In all experiments, we firstly asked all the participants to read the definition of the healthcare chatbot to enrich their knowledge of it. Then, to better manipulate the healthcare chatbot service failure, we used different situation statements as stimulus material for each experiment. 4 EXPERIMENT 1: MAIN EFFECT OF ROBOT SERVICE-PATIENT TRUST 4.1 EXPERIMENTAL DESIGN Experiment 1 aimed to explore the main effect of robot services on patient trust. We designed a one-way between-subjects design (robot service: process failure vs. outcome failure), and recruited 300 participants on credamo ( https://www.credamo.com/ ), of which 147 (49.0%) were male and 153 (51.0%) were female. The age distribution of all the participants was 24.0% aged 18–25 years, 30.0% aged 26–40 years, 24.3% aged 41–60 years and 21.7% aged 61 years and above. All participants were randomly divided into two groups and assigned to two different experimental scenarios (outcome failure group Vs. process failure group). The process failure group with a total of 150 participants got the corresponding answers although questions arose during the process; the outcome failure group with a total of 150 participants failed to get the corresponding answers because the questions arose during the process. 4.2 EXPERIMENTAL PROCESS First, all participants were guided to imagine that they were in a medical institution, and after typing in a medical question to the healthcare chatbot, they took a long time and repeatedly pressed the "Answer Output" button on the screen. Participants in the process failure group got the answer to the question after several attempts. However, the participant in the outcome failure group got the hospital's information after several attempts. Next, we asked the participants, "Do you agree that you were satisfied with your interaction with the healthcare chatbot?", "Do you agree that you developed a good feeling about the healthcare chatbot? ", "Do you agree that you felt a strong sense of positive engagement?", and "Do you agree that in interacting with the healthcare chatbot, it played an influential interactive role for the patient?" (1 = Strongly Disagree, 7 = Strongly Agree) to measure participants' trust. The scale was adapted from the Trust Scale developed by Jin and Eastin ( 2022 ). Based on the study of Watson et al. ( 1988 ), we found that the emotional state of the patient has a significant impact on the patient trust, and in order to enhance the accuracy of the experimental results, we need to manipulate the emotions of the participants. We asked the participants "Do you have a good emotion and are happy, pleased, and satisfied?" (1 = Strongly Disagree, 7 = Strongly Agree), a question taken from the scale of Watson et al. ( 1988 ). Last, we collected demographic information about the participants (Cronbach's α = 0.655). 4.3 EXPERIMENTAL RESULTS There was a main effect test. We employed one-way ANOVA with healthcare chatbot service failure as the independent variable and patient trust as the dependent variable. The results showed that patient trust in the robot service process failure group (M = 5.02, SD = 1.002) was significantly higher than that in the outcome failure group (M = 3.55, SD = 0.843), F (1, 298) = 189.754, p < 0.001. H1 was verified. There was a control variable analysis. Based on the study of Pollitt ( 1982 ), we found that emotion is an important factor that could affect individual trust. Therefore, we conducted an analysis of covariance (ANCOVA) with patient trust as the dependent variable, healthcare chatbot service as the independent variable, and patients' emotions as the covariate. The results of the experiment showed that patient emotion did not have a significant effect on patient trust (F (1, 298) = 28.308, p < 0.001). Therefore, patient emotion did not affect the results of the experiment, and H1 was again verified. 4.4 DISCUSSION Experiment 1 verified that healthcare chatbot service has a significant positive impact on patient trust, and the process failure of healthcare chatbot service affects patient trust more than the outcome failure. We also verified that there was no significant effect of patient emotions on patient trust so as to enhance the reliability and accuracy of the experiment. Despite the above findings, Experiment 1 failed to further explore the internal mechanisms and boundary conditions between the healthcare chatbot and patient trust. Experiment 2 introduced patient mindset as a moderating variable to explore the interaction effect between patient mindset and robot service failure on patient trust. 5 EXPERIMENT 2: INTERACTION EFFECT BETWEEN PATIENT MINDSET AND HEALTHCARE CHATBOT SERVICE FAILURE 5.1 EXPERIMENTAL DESIGN Experiment 2 aimed mainly at: first, re-validating the main effect; and second, exploring the interaction effect between patient mindset (growth vs. fixed) and healthcare chatbot service failure on patient trust. We designed a 2 (robot service: process failure vs. outcome failure) X 2 (patient mindset: growth vs. fixed) ANOVA and recruited 600 participants on credamo ( https://www.credamo.com/ ), of which 301 (50.2%) were male and 299 (49.8%) were female. The age distribution of all the participants was 24.5% aged 18–25 years, 27.8% aged 26–40 years, 25.5% aged 41–60 years and 22.2% aged 60 years and above. All participants were then assigned to two different experimental scenarios (outcome failure group Vs. process failure group). The 302 participants in the process failure group got the corresponding answers although the questions arose during the process. In the process failure group, we used the scenario material to subdivide them into the process failure-growth group (N = 153) and the process failure-fixed group (N = 149) with the patient's mindset. The 298 participants in the outcome failure group failed to get the corresponding answers because the questions arose during the process. In the outcome failure group, we used the scenario material to subdivide them into the outcome failure-growth group (N = 148) and the outcome failure-Fixed group (N = 150). 5.2 EXPERIMENTAL PROCESS First, all participants were guided to imagine that they were in a medical institution, and after typing in a medical question to the healthcare chatbot, they took a long time and repeatedly pressed the "Answer Output" button on the screen. Participants in the process failure group got the answer to the question after several attempts. However, the participant in the outcome failure group got the hospital's information after several attempts. We all told the participants of each fixed group that the programming of the healthcare chatbot can be continuously learnt and improved, but all robots have limitations in their growth, and its service capability will be fixed for and will not grow a long time to come; while we told the participants of each growth group that the programming of the robot can be continuously learnt and improved, and the service and communication methods are gradually humanized and scientized, and are able to satisfy personal companionship (“need”). Second, we asked the participants in the fixed group, "Do you agree that no matter how capable the healthcare chatbot is, we can always change it more and help it grow?", and "Do you agree that we can always change the healthcare chatbot's capabilities significantly?" (1 = Strongly Disagree, 7 = Strongly Agree); and asked participants in the growth group, "Do you agree that the healthcare chatbot has some limits to its capabilities, that you can't ask too much of it, and that its evolution is slow?", and "Do you agree that a healthcare chatbot can learn new things, but we can't really change the basic level of competence of a healthcare chatbot?" (1 = strongly disagree, 7 = strongly agree) Patient mindset was measured using two different sets of questions adapted from the scale of Chiu et al. ( 1997 ). Next, we asked the participants, "Do you agree that you were satisfied with your interaction with the healthcare chatbot?", "Do you agree that you developed a good feeling about the healthcare chatbot?", "Do you agree that you felt a strong sense of positive engagement?", and "Do you agree that in interacting with the healthcare chatbot, it played an influential interactive role for the patient?" (1 = Strongly Disagree, 7 = Strongly Agree) to measure participants' trust. The scale was adapted from the Trust Scale developed by Jin and Eastin ( 2022 ). Last, we collected demographic information about the participants (Cronbach's α = 0.68). 5.3 EXPERIMENTAL RESULTS There was a main effect test. We conducted a one-way ANOVA with healthcare chatbot service failure as the independent variable and patient trust as the dependent variable. The experimental results showed that patient trust in process failure group (M = 4.36, SD = 1.062) was significantly higher than that in outcome failure group (M = 4.1527, SD = 1.272), F (1, 598) = 4.702, p = 0.031. H1 was verified. There was a manipulation check. We conducted a one-way ANOVA with patient mindset as the independent variable and patient trust as the dependent variable. The results of the experiment showed that M growth group = 4.931, SD growth group = 1.018; M fixed group = 3.578, SD fixed group = 0.9; and F (1, 598) = 296.815, p < 0.001. It can be seen that the patient trust in the growth mindset group was significantly higher than that in the fixed mindset group. Therefore, the manipulation in Experiment 2 was successful. There was an interaction effect test. We conducted a 2 (robot service: process failure vs. outcome failure) X 2 (patient mindset: growth vs. fixed) ANOVA with patient trust as the dependent variable, in an attempt to validate the interaction effect between patient mindset and healthcare chatbot service failure on patient trust. The results showed that healthcare chatbot service failure had a significant effect on patient trust (F (1, 598) = 6.355, p < 0.012), patient mindset had a significant effect on patient trust (F (1, 598) = 306.081, p < 0.001), and the interaction between patient mindset and healthcare chatbot service failure had a significant effect on patient trust (F (1, 598) = 14.458, P < 0.001). It can be seen that the interaction between patient mindset and healthcare chatbot service failure had a significant effect on patient trust, which validated H2.(shown in Fig. 1 ). [Insert Fig. 1 about here] There was a control variable analysis. We considered that gender may affect patient trust after healthcare chatbot service failure Barone et al. ( 2024 ). Therefore, we employed one-way ANOVA with gender as the independent variable and patient trust as the dependent variable. The results showed that patient gender had no significant effect on patient trust (F (1,598) = 0.022, p = 0.882). Therefore, patient gender did not affect the results of the experiment and H1 was again validated. 5.4 DISCUSSION Experiment 2 verified the interaction effect between patient mindset and healthcare chatbot service failure on patient trust. The experimental results showed that in the face of healthcare chatbot service failure, patient trust in the growth mindset group was significantly higher than that in the fixed mindset group. Specifically, with the healthcare chatbot service process failure, the difference between the patient trust in the growth mindset group and the fixed mindset group is smaller, showing a tendency to intersect; with the healthcare chatbot service outcome failure, the difference between the patient trust in the growth mindset group and the fixed group is larger, showing a tendency to diverge. At the same time, we excluded the impact of patients' gender on the experimental results, which enhanced the accuracy of the experiment. Despite the above findings in Experiment 2, the impact of the potential factors from the healthcare chatbot service itself on patient trust has not been further explored. Therefore, in order to address the above research gap, Experiment 3 introduced a healthcare chatbot linguistic expression type to investigate its interaction effect with healthcare chatbot service failure on patient trust. 6 EXPERIMENT 3: INTERACTION EFFECT BETWEEN LINGUISTIC EXPRESSION TYPE AND HEALTHCARE CHATBOT SERVICE FAILURE 6.1 EXPERIMENTAL DESIGN Experiment 3 aimed to investigate the interaction effect of linguistic expression type and healthcare chatbot service failure on patient trust. We designed 2(healthcare chatbot service failure: process failure vs. outcome failure) X 2(linguistic expression type: need vs. want) ANOVA and recruited 520 participants on credamo ( https://www.credamo.com/ ), of which 259 (49.8%) subjects were male and 261 (50.2%) subjects were female. The age distribution of all participants was 20% aged 18–25 years, 31.5% aged 26–40 years, 25.2% aged 41–60 years and 23.3% aged 60 years and above. All participants were then assigned to two different experimental scenarios (outcome failure group Vs. process failure group). The participants in the process failure group got the corresponding answers although the questions arose during the process. In the process failure group, we used scenario materials of linguistic expression type to subdivide participants into process failure-need group (N = 94) and process failure-want group (N = 146). The participants in the outcome failure group failed to get the corresponding answers because the questions arose during the process. In the outcome failure group, we used the scenario materials of linguistic expression type to subdivide the participants into process failure-need group (N = 150) and process failure-want group (N = 130). 6.2 EXPERIMENTAL PROCESS First, all participants were guided to imagine that they were in a medical institution, and after typing in a medical question to the healthcare chatbot, they took a long time and repeatedly pressed the "Answer Output" button on the screen. Participants in the process failure group got the answer to the question after several attempts. However, the participant in the outcome failure group got the hospital's information after several attempts. Then, in the need group, the healthcare chatbot asked the participant: "I'm sorry, I made a mistake in my understanding, I need you to retype the question, can we proceed with the diagnosis?"; and in the want group, the healthcare chatbot asked the participant: "I'm sorry, I made a mistake in my understanding, I want to ask for your understanding, can you retype the question, and let's proceed with the diagnosis?". Then, we asked the participants, "Do you agree that you were satisfied with your interaction with the healthcare chatbot?", "Do you agree that you developed a good feeling about the healthcare chatbot?", "Do you agree that you felt a strong sense of positive engagement?", and "Do you agree that in interacting with the healthcare chatbot, it played an influential interactive role for the patient?" (1 = Strongly Disagree, 7 = Strongly Agree) to measure participants' trust. The scale was adapted from the Trust Scale developed by Jin and Eastin ( 2022 ). Based on the study of Kamide et al. ( 2014 ), we found that psychological acceptance of the robot may have an impact on patient trust. Therefore, to manipulate patients' psychological acceptance, we asked participants "Do you agree that the robot made you feel uncomfortable?" and "Do you agree that robots seem to reduce the value of the human healthcare workers?" (1 = Strongly Disagree, 7 = Strongly Agree), which was adapted from the Psychological Acceptance Scale developed by Kamide et al. ( 2014 ). Last, we collected demographic information about the participants (Cronbach's α = 0.634). 6.3 EXPERIMENTAL RESULTS There was a main effect test. We conducted a one-way ANOVA with healthcare chatbot service failure (process failure Vs. outcome failure) as the independent variable and patient trust as the dependent variable. The experimental results showed that patient trust in service process failure group (M = 4.511, SD = 1.041) was significantly higher than that in service outcome failure group (M = 3.957, SD = 1.291), F (1, 518) = 28.375, p < 0.001. H1 was validated. There was a manipulation check. We conducted a one-way ANOVA with the linguistic expression types (need vs. want) as the independent variable and patient trust as the dependent variable. The results of the experiment showed that M need =4.956, SD need =0.902; M want =3.371, SD want =0.941; F (1, 518) = 383.344, p < 0.001. It can be seen that patient trust in the want group was significantly higher than in the need group, and the manipulation in experiment 3 was successful. There was an interaction effect test. We analyzed the interaction effect of healthcare chatbot service failure and robot linguistic expression type, with patient trust as the dependent variable, by performing a 2 (healthcare chatbot service failure: process failure vs. outcome failure) X 2 (linguistic expression type: need vs. want) ANOVA. The results showed that healthcare chatbot service failure had a significant effect on patient trust (F (1, 518) = 19.95, p < 0.001), and the linguistic expression type of the healthcare chatbot had a significant effect on patient trust (F (1, 518) = 355.31, p < 0.001), and the interaction between the linguistic expression type of the healthcare chatbot and the healthcare chatbot service failure showed a significant effect on patient trust (F (1, 518) = 12.535, p < 0.001). It can be seen that the interaction effect between linguistic expression type and healthcare chatbot service failure was significant on patient trust, validating H3.(shown in Fig. 2 ). [Insert Fig. 2 about here] There was a control variable analysis. Based on the study by Kamide et al. ( 2014 ), we found that patient acceptance to robot may have an impact on patient trust. Therefore, we used patient acceptance as the independent variable and patient trust as the dependent variable, and the results of the experiment showed that patient acceptance did not have a significant effect on patient trust (F (12, 507) = 0.745, p = 0.707). 6.4 DISCUSSION Experiment 3 demonstrated the interaction effect between linguistic expression type and healthcare chatbot service failure on patient trust. The main effect of healthcare chatbot service failure on patient trust was again validated. The results of the experiment pointed out that faced the robot service failure, patient trust was significantly higher in the “need” group than in the “want” group. Specifically, with the healthcare chatbot service process failure, the difference between patient trust in the “need” group and the “want” group is smaller, showing a tendency to intersect; with the healthcare chatbot service outcome failure, the difference between the two is larger, showing a tendency to diverge. At the same time, we exclude the impact of the patient acceptance toward the robot on the experimental results, which enhance its robustness. 7 GENERAL DISCUSSION What kind of message should be sent to conveys important information in a healthcare chatbot service failure scenario. Research has generally found that healthcare chatbot service failure significantly reduces patient trust (Park, 2020 ). However, this study showed that healthcare chatbot attempts to redeem patient trust with messages that help patients re-establish their trust in the robot. While sending out messages is indeed an integral part of successfully building human-robot trust, academics have not yet considered why messages sent by a healthcare chatbot can reshape trust between humans and robots. This study addressed a question that has not been adequately investigated: how robots affect patient trust with service failure. We proposed and tested an idea that different semantic frames (need vs. want) in sentences sent by the robot in a healthcare chatbot service failure (process failure vs. outcome failure) scenario affects patients' evaluations for the value of the healthcare chatbot and their subsequent willingness to trust. More specifically, we focused on the significance of patient mindset (growth vs. fixation) and semantic frames (need vs. want) in the healthcare chatbot service failure scenario to enhance patient trust (Su et al., 2024 ). The experiment demonstrated that patients do associate healthcare chatbot service failure scenarios with their own mindset, especially when confronted with a robot service process failure. We further showed why the healthcare chatbot semantic framework (need vs. want) can be effective in restoring patients and ultimately enhancing patient trust. This study also indicated the meaningful impact of patient mindset (growth vs. fixed), and robot semantic frames (need vs. want) on regaining patient trust in the face of healthcare chatbot service failure. We suggested that in the case of healthcare chatbot service failure, sending more sentences with "want" would increase patient trust; conversely, sending sentences with "need" would decrease patient trust. However, there was a significant interaction between healthcare chatbot service failure (process failure vs. outcome failure) and patient mindset (growth vs. fixed). Th pattern of this effect adds complexity to the theory of this study. When a healthcare chatbot interacts with patients, they are more likely to be influenced by their own attitudes toward the robot, which ultimately affects patient trust in a healthcare chatbot service failure scenario. 7.1 THEORETICAL IMPLICATIONS An important contribution that this study has made is to the literature on robot services and patient psychological behaviors, by highlighting that the sent sentences in healthcare chatbot service failure scenarios with different semantic frames, can affect patients' evaluations of the healthcare chatbot and generate different patient trusts. Despite the fact that service robots are increasingly used to improve individual trust in it through verbal interactions (Tussyadiah et al., 2020 ). A large number of studies suggested that the scenarios of robot service failure can salvage individual satisfaction in multiple ways (Belanche et al., 2020 ; Liao et al., 2023 ). However, academics have rarely considered the important mechanism of robots influencing patient trust in healthcare chatbot service failure scenarios. As a notable exception, Choi et al. ( 2021 ) examined efforts made after a healthcare chatbot service failure, but they did not consider the semantic frames in the sent sentences (need vs. want) as well as the patient mindset (growth vs. fixed). Our goal is to bridge the gap between practice and academic research by introducing robust theoretical and experimental approaches to help understand whether patient mindset and semantic frames matter for healthcare chatbot service failure. Our work showed that the question of why and when a healthcare chatbot is able to regain patient trust in the service failure, contributing to the advancement of knowledge on healthcare chatbot service and patient psychological behaviors. Thus, our study unlocked many pathways for academics to learn more about how patient trust can be regained with robot service failure. This study also contributed to the development of semantic frames, by integrating semantic frames (need vs. want) and revealing the key mechanisms by which different sent semantic sentences in healthcare chatbot service failure scenarios affect patient trust (deRosset, 2024 ). So far, research on semantic frames has only indirectly explored underlying psychological mechanisms (Bless et al., 1998 ; P. X. Wang et al., 2023 ) and has mainly relied on general semantic symbols to perceive scenarios (Beuls et al., 2021 ). Therefore, this study has helped to understand the information effect of semantic frame in robot service scenarios through in-depth discussions. In addition to extending the scenarios of healthcare chatbot service failure affecting patient trust to everyday services, the theoretical and empirical research also adds to the understanding of patient mindset. As other literature has shown, people's mindset also determines how they perceive robots (Dang & Liu, 2022 ). The rigorously controlled real-life experiments in this study are equally important in supporting the impact of patient mindset on patient trust with robot service failure. Based on previous literature on robot service failure, this study demonstrated multiple ways in which service failure can be altered. The experimental results showed that the mechanisms and principles of both process failure and outcome failure in healthcare chatbot service are related to the personal attributes of the patient. These findings can help us to reposition robot service failure from external and uncontrollable factors. Previous studies have found that the factors affecting individual trust lie more in the control of the individual themselves, which helps to redirect the topic of research on robot service failure away from external and uncontrollable factors towards factors of individuals' own control. Although theorists Leo and Huh ( 2020 ) have hypothesized that compensatory actions in service robot failure scenarios are effective in driving individuals' willingness to adopt service robots again. However, these ideas have so far remained untested by substantial empirical evidence. This study can also address the question of how to regain individuals' trust in healthcare chatbot with service failure. 7.2 PRACTICAL IMPLICATIONS This study demonstrated that the initiative of service robot in sending seemingly irrelevant messages may be beneficial for regaining patients' trust. It is worth noting that we do not believe that the specific discourse of service robots can be the most important means to regain patient trust. However, a relatively simple and feasible discourse for service robots can help to rebuild patient trust in healthcare chatbot service failure scenarios, which is important for healthcare chatbot in real-life activities. Of course, robot discourse is not an intervention in the traditional sense, but something that every healthcare chatbot maker and user should consider. Therefore, this study indicated that an additional set of "need" statements for the healthcare chatbot could better compensate for actual healthcare chatbot service failure scenarios, which could have a significant impact on the performance of service robots in the medical institutes. In addition, the magnitude of the observed effect sizes can effectively affect the practical implications of the study through its design and variables. They can have a smaller impact on purely scenario-based experimental studies that would normally have a larger significance in practice (Bakan, 1966 [Mohajeri, 2020 #96)]. The patient trust as an explanatory variable implies that patients' daily healthcare activities usually require contact with either a doctor or a healthcare chatbot, indicating the need and urgency for healthcare chatbot service improvement. Therefore, by exploring the types of healthcare chatbot service failure and how the words sent by the healthcare chatbot in these scenarios affect patient trust, we could provide more practical suggestion. Different information can deal with different scenarios, so effectively designing multiple semantic frames in a single human-robot interaction is crucial for users of healthcare chatbots. Past research has demonstrated that semantic frames are more effective in engaging as well as changing the behavioral intentions of individuals (Gao & Zhang, 2021 ), and semantic frames for healthcare chatbots provide great service value to medical institutes (Taylor, 2006 ). Finding a way to combine the semantic frames and service failure scenarios in the design of a healthcare chatbot is clearly beneficial to both the manufacturer and the user of the healthcare chatbot (medical institutes and patients). The findings suggested that healthcare chatbot manufacturers should adopt different strategies for dealing with healthcare chatbot service failures. For example, when patients facing the process failures, it is advisable to encourage the patients to re-ask the question in a direct tone. When patients facing the outcome failures, it is advisable to acknowledge the occurred mistake and issue a new statement to help the patient to re-ask their questions. Similarly, in healthcare chatbot service failure scenarios, considering the patient mindset and regaining the patient trust can be an effective service tool. However, if healthcare chatbot makers want to regain patient trust by encouraging patient tolerance towards the healthcare chatbot, they should consider guiding patients in advance. For example, informing patients that healthcare chatbots are still in a developmental stage and that they need patients' tolerance and patience. Additionally, creating a growth mindset for patients towards healthcare chatbots is also crucial if medical institutes want to improve the scenarios of healthcare chatbot service failure. 7.3 LIMITATIONS AND FUTURE RESEARCH Although this study provided new insights into patient psychological behaviors and robot service failure, as well as how they can be used to regain patient trust, there are some limitations that need to be addressed in future research. First, we manipulated the patient mindset, but an inclusive growth mindset towards novelty that is widespread in the human world due to the specificity of individual patients. Consequently, the growth mindset may also lead patients to believe that healthcare chatbot service failures are inevitable, as patients with this mindset may believe that all healthcare chatbots need to be tolerated. Future research could consider a more appropriate experimental design to avoid the confounding impact of the growth mindset on individual patients. Second, the manipulations of healthcare chatbot service failures were mainly classified it into process failure and outcome failure. However, in real-life scenario, healthcare chatbot service failures may include more contingencies, such as service lag or delay. Future research could perhaps manipulate service failures to explore the diverse antecedents of robot service failure to restore individual trust. Last, we argued previously that the language sent by the healthcare chatbot in service failure would be directly effective in affecting patients' subsequent behavior. However, the role of an alternative explanation - patients' skepticism towards the robot's language - has not been explored in this study and could be further investigated in future research. Declarations Compliance with Ethical Standards Conflict of interest The authors declare that they have no conflict of interest. Funding The study received no funding from any organization or individual. Method All methods were carried out in accordance with relevant guidelines and regulations. All experimental protocols were approved by a named institutional and/or licensing committee. Informed consent was obtained from all subjects and/or their legal guardian(s) This study was reviewed and approved by the Ethics Committee of Huainan Normal University. References Alhaffar, M. B. A., & Janos, S. (2021). Public health consequences after ten years of the Syrian crisis: a literature review. Globalization and health , 17 (1), 111. https://doi.org/https://doi.org/10.1186/s12992-021-00762-9 Arikan, E., Altinigne, N., Kuzgun, E., & Okan, M. (2023). May robots be held responsible for service failure and recovery? The role of robot service provider agents’ human-likeness. Journal of Retailing and Consumer Services , 70 , 103175. https://doi.org/https://doi.org/10.1016/j.jretconser.2022.103175 Asan, O., Bayrak, A. E., & Choudhury, A. (2020). Artificial intelligence and human trust in healthcare: focus on clinicians. Journal of Medical Internet Research , 22 (6), e15154. https://doi.org/https://doi.org/10.2196/15154 Bakan, D. (1966). The test of significance in psychological research. Psychological bulletin , 66 (6), 423. https://doi.org/https://doi.org/10.1037/h0020412 Baker, D. E. (2018). How Want vs Need Self-Talk Facilitates Goal-Directed Behavior . University of Arkansas. Barone, A. M., Stagno, E., & Donato, C. (2024). Call it robot: anthropomorphic framing and failure of self-service technologies. Journal of Services Marketing , 38 (3), 272-287. https://doi.org/https://doi.org/10.1108/JSM-05-2023-0169 Becker, M., Mahr, D., & Odekerken-Schröder, G. (2023). Customer comfort during service robot interactions. Service Business , 17 (1), 137-165. https://doi.org/https://doi.org/10.1007/s11628-022-00499-4 Belanche, D., Casaló, L. V., Flavián, C., & Schepers, J. (2020). Service robot implementation: a theoretical framework and research agenda. The Service Industries Journal , 40 (3-4), 203-225. https://doi.org/https://doi.org/10.1080/02642069.2019.1672666 Beuls, K., Van Eecke, P., & Cangalovic, V. S. (2021). A computational construction grammar approach to semantic frame extraction. Linguistics Vanguard , 7 (1), 20180015. https://doi.org/https://doi.org/10.1515/lingvan-2018-0015 Bless, H., Betsch, T., & Franzen, A. (1998). Framing the framing effect: The impact of context cues on solutions to the ‘Asian disease’problem. European journal of social psychology , 28 (2), 287-291. https://doi.org/https://doi.org/10.1002/(SICI)1099-0992(199803/04)28:23.0.CO;2-U Bock, D. E., Wolter, J. S., & Ferrell, O. (2020). Artificial intelligence: Disrupting what we know about services. Journal of Services Marketing , 34 (3), 317-334. https://doi.org/https://doi.org/10.1108/JSM-01-2019-0047 Brooks, R. A. (1991). New approaches to robotics. Science , 253 (5025), 1227-1232. https://doi.org/https://doi.org/10.1126/science.253.5025.1227 Chao, S., Mu, Q. S., Austin, J., & Mueller, C. M. (2023). Parental Health Mindset Is Related to Preoperative Perceptions of Chest Wall Anomalies. Journal of Surgical Research , 284 , 318-321. https://doi.org/https://doi.org/10.1016/j.jss.2022.11.046 Chi, O. H., Chi, C. G., Gursoy, D., & Nunkoo, R. (2023). Customers’ acceptance of artificially intelligent service robots: The influence of trust and culture. International Journal of Information Management , 70 , 102623. https://doi.org/https://doi.org/10.1016/j.ijinfomgt.2023.102623 Ching, B. C., Bennett, S. D., Morant, N., Heyman, I., Schleider, J. L., Fifield, K., Allen, S., & Shafran, R. (2023). Growth mindset in young people awaiting treatment in a paediatric mental health service: A mixed methods pilot of a digital single-session intervention. Clinical Child Psychology and Psychiatry , 28 (2), 637-653. https://doi.org/https://doi.org/10.1177/13591045221105193 Chiu, C.-y., Dweck, C. S., Tong, J. Y.-y., & Fu, J. H.-y. (1997). Implicit theories and conceptions of morality. Journal of personality and social psychology , 73 (5), 923. https://doi.org/ https://doi.org/10.1037/0022-3514.73.5.923 Choi, S., Mattila, A. S., & Bolton, L. E. (2021). To err is human (-oid): how do consumers react to robot service failure and recovery? Journal of Service Research , 24 (3), 354-371. https://doi.org/https://doi.org/10.1177/1094670520978798 Chuah, S. H.-W., Aw, E. C.-X., & Yee, D. (2021). Unveiling the complexity of consumers’ intention to use service robots: An fsQCA approach. Computers in Human Behavior , 123 , 106870. Cowger, J. A., Grafton, G., & Shah, P. (2020). Avoiding the “set it and forget it mentality”: a need to regularly reassess left ventricular assist device patients for optimal support. The Journal of thoracic and cardiovascular surgery , 159 (4), 1322-1325. https://doi.org/https://doi.org/10.1016/j.jtcvs.2019.06.134 Dang, J., & Liu, L. (2022). A growth mindset about human minds promotes positive responses to intelligent technology. Cognition , 220 , 104985. https://doi.org/https://doi.org/10.1016/j.cognition.2021.104985 deRosset, L. (2024). A Semantic Framework for the Impure Logic of Ground. Journal of Philosophical Logic , 1-29. https://doi.org/https://doi.org/10.1007/s10992-023-09738-z Desveaux, L., & Ivers, N. (2024). Practice or perfect? Coaching for a growth mindset to improve the quality of healthcare. BMJ Quality & Safety . https://doi.org/https://doi.org/10.1136/bmjqs-2023-016456 Doupe, P., Faghmous, J., & Basu, S. (2019). Machine learning for health services researchers. Value in Health , 22 (7), 808-815. https://doi.org/https://doi.org/10.1016/j.jval.2019.02.012 Dupont, P. E., Nelson, B. J., Goldfarb, M., Hannaford, B., Menciassi, A., O’Malley, M. K., Simaan, N., Valdastri, P., & Yang, G.-Z. (2021). A decade retrospective of medical robotics research from 2010 to 2020. Science robotics , 6 (60), eabi8017. https://doi.org/https://doi.org/10.1126/scirobotics.abi8017 Fissell, R. B., Schlundt, D., Cavanaugh, K. L., Mueller, C., Umeukeje, E. M., Nair, D., Wild, M., Chariwala, S., Guide, A., & Stewart, T. (2021). A cross-sectional study of fixed and growth mindset in adult peritoneal dialysis patients. Journal of Patient Experience , 8 , 23743735211055289. https://doi.org/https://doi.org/10.1177/23743735211055289 Foster, M. E. (2019). Natural language generation for social robotics: opportunities and challenges. Philosophical Transactions of the Royal Society B , 374 (1771), 20180027. https://doi.org/https://doi.org/10.1098/rstb.2018.0027 Gao, Y., & Zhang, Z. (2021). Consensus reaching with non-cooperative behavior management for personalized individual semantics-based social network group decision making. Journal of the Operational Research Society , 73 (11), 2518-2535. https://doi.org/https://doi.org/10.1080/01605682.2021.1997654 Gibelli, F., Ricci, G., Sirignano, A., Turrina, S., & De Leo, D. (2021). The increasing centrality of robotic technology in the context of nursing care: bioethical implications analyzed through a scoping review approach. Journal of Healthcare Engineering , 2021 . https://doi.org/https://doi.org/10.1155/2021/1478025 Grant, A. A. (2023). Regaining the student mindset of patient care. Annals of Vascular Surgery . https://doi.org/https://doi.org/10.1016/j.avsg.2023.12.009 Guo, C., & Li, H. (2022). Application of 5G network combined with AI robots in personalized nursing in China: A literature review. Frontiers in Public Health , 10 , 948303. https://doi.org/https://doi.org/10.3389/fpubh.2022.948303 Guo, Y., Yang, Y., Liu, Y., Li, Q., Cao, F., Feng, M., Wu, H., Li, W., & Kang, Y. (2021). Development status and multilevel classification strategy of medical robots. Electronics , 10 (11), 1278. https://doi.org/https://doi.org/10.3390/electronics10111278 Han, J., Davids, J., Ashrafian, H., Darzi, A., Elson, D. S., & Sodergren, M. (2022). A systematic review of robotic surgery: From supervised paradigms to fully autonomous robotic approaches. The International Journal of Medical Robotics and Computer Assisted Surgery , 18 (2), e2358. https://doi.org/ https://doi.org/10.1002/rcs.2358 Holland, J., Kingston, L., McCarthy, C., Armstrong, E., O’Dwyer, P., Merz, F., & McConnell, M. (2021). Service robots in the healthcare sector. Robotics , 10 (1), 47. https://doi.org/ https://doi.org/10.3390/robotics10010047 Holthöwer, J., & van Doorn, J. (2023). Robots do not judge: service robots can alleviate embarrassment in service encounters. Journal of the Academy of Marketing Science , 51 (4), 767-784. Holzinger, A., Langs, G., Denk, H., Zatloukal, K., & Müller, H. (2019). Causability and explainability of artificial intelligence in medicine. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery , 9 (4), e1312. https://doi.org/ https://doi.org/10.1002/widm.1312 Hu, Y., Min, H., & Su, N. (2021). How sincere is an apology? Recovery satisfaction in a robot service failure context. Journal of Hospitality & Tourism Research , 45 (6), 1022-1043. https://doi.org/https://doi.org/10.1177/10963480211011533 Huang, R., Li, H., Suomi, R., Li, C., & Peltoniemi, T. (2023). Intelligent physical robots in health care: systematic literature review. Journal of Medical Internet Research , 25 , e39786. https://doi.org/https://doi.org/10.2196/39786 Hwang, J., Park, S., & Kim, I. (2020). Understanding motivated consumer innovativeness in the context of a robotic restaurant: The moderating role of product knowledge. Journal of Hospitality and Tourism Management , 44 , 272-282. Jia, Y., Wyer Jr, R. S., & Shen, H. (2021). “Will you?” versus “can you?”: Verbal framing moderates the effect of feelings of power on consumers’ reactions to waiting. Journal of Experimental Psychology: Applied , 27 (2), 213. https://doi.org/https://doi.org/10.1037/xap0000336 Jin, E., & Eastin, M. S. (2022). When a chatbot smiles at you: The psychological mechanism underlying the effects of friendly language use by product recommendation chatbots. Cyberpsychology, Behavior, and Social Networking , 25 (9), 597-604. https://doi.org/https://doi.org/10.1089/cyber.2021.0318 Joseph, G. P., Segovia, N. A., Wright, R. C., Mueller, C., & Tileston, K. R. (2021). Mindset correlates with health-related quality of life assessment in patients with adolescent idiopathic scoliosis. Spine Deformity , 9 , 349-354. https://doi.org/https://doi.org/10.1007/s43390-020-00243-w Kakhi, K., Alizadehsani, R., Kabir, H. D., Khosravi, A., Nahavandi, S., & Acharya, U. R. (2022). The internet of medical things and artificial intelligence: trends, challenges, and opportunities. Biocybernetics and Biomedical Engineering , 42 (3), 749-771. https://doi.org/https://doi.org/10.1016/j.bbe.2022.05.008 Kamide, H., Kawabe, K., Shigemi, S., & Arai, T. (2014). Relationship between familiarity and humanness of robots–quantification of psychological impressions toward humanoid robots. Advanced Robotics , 28 (12), 821-832. https://doi.org/https://doi.org/10.1080/01691864.2014.893837 Kerasidou, A. (2020). Artificial intelligence and the ongoing need for empathy, compassion and trust in healthcare. Bulletin of the World Health Organization , 98 (4), 245. https://doi.org/https://doi.org/10.2471/BLT.19.237198 Khaksar, S. M. S., Shahmehr, F. S., Miah, S., Daim, T., & Ozdemir, D. (2024). Privacy concerns versus personalisation benefits in social robot acceptance by employees: A paradox theory—Contingency perspective. Technological Forecasting and Social Change , 198 , 123034. https://doi.org/https://doi.org/10.1016/j.techfore.2023.123034 Koutentakis, D., Pilozzi, A., & Huang, X. (2020). Designing socially assistive robots for Alzheimer’s disease and related dementia patients and their caregivers: where we are and where we are headed. Healthcare, Kuchenbrandt, D., Eyssel, F., Bobinger, S., & Neufeld, M. (2013). When a robot’s group membership matters: Anthropomorphization of robots as a function of social categorization. International Journal of Social Robotics , 5 , 409-417. https://doi.org/https://doi.org/10.1007/s12369-013-0197-8 Lee, H., Piao, M., Lee, J., Byun, A., & Kim, J. (2020). The purpose of bedside robots: exploring the needs of inpatients and healthcare professionals. CIN: Computers, Informatics, Nursing , 38 (1), 8-17. https://doi.org/https://doi.org/10.1097/CIN.0000000000000558 Leo, X., & Huh, Y. E. (2020). Who gets the blame for service failures? Attribution of responsibility toward robot versus human service providers and service firms. Computers in Human Behavior , 113 , 106520. https://doi.org/https://doi.org/10.1016/j.chb.2020.106520 Liao, G.-Y., Huang, T.-L., Wong, M.-K., Shyu, Y.-I. L., Ho, L.-H., Wang, C., Cheng, T., & Teng, C.-I. (2023). Enhancing Nurse–Robot Engagement: Two-Wave Survey Study. Journal of Medical Internet Research , 25 , e37731. https://doi.org/https://doi.org/10.2196/37731 Liu, D., Li, C., Zhang, J., & Huang, W. (2023). Robot service failure and recovery: Literature review and future directions. International Journal of Advanced Robotic Systems , 20 (4), 17298806231191606. https://doi.org/https://doi.org/10.1177/17298806231191606 Liu, J., Kong, X., Xia, F., Bai, X., Wang, L., Qing, Q., & Lee, I. (2018). Artificial intelligence in the 21st century. Ieee Access , 6 , 34403-34421. https://doi.org/https://doi.org/10.1109/ACCESS.2018.2819688 Liu, X., He, X., Wang, M., & Shen, H. (2022). What influences patients' continuance intention to use AI-powered service robots at hospitals? The role of individual characteristics. Technology in Society , 70 , 101996. https://doi.org/https://doi.org/10.1016/j.techsoc.2022.101996 Luo, Z., Cheng, W., Zhao, T., & Xiang, N. (2024). Intelligent sensory systems toward soft robotics. Applied Materials Today , 37 , 102122. https://doi.org/https://doi.org/10.1016/j.apmt.2024.102122 Mende, M., Scott, M. L., Van Doorn, J., Grewal, D., & Shanks, I. (2019). Service robots rising: How humanoid robots influence service experiences and elicit compensatory consumer responses. Journal of Marketing Research , 56 (4), 535-556. https://doi.org/https://doi.org/10.1177/0022243718822827 Meng, Y.-J., Li, N.-X., Liu, C.-J., Chen, J.-H., Song, Y.-C., & Qian, Z.-S. (2008). Quality of life and hostile mentality trend of patients with HIV/AIDS in China. Public Health , 122 (4), 404-411. https://doi.org/https://doi.org/10.1016/j.puhe.2007.06.007 Moriuchi, E., & Murdy, S. (2024). The role of robots in the service industry: Factors affecting human-robot interactions. International Journal of Hospitality Management , 118 , 103682. https://doi.org/https://doi.org/10.1016/j.ijhm.2023.103682 Packard, G., & Berger, J. (2017). How language shapes word of mouth's impact. Journal of Marketing Research , 54 (4), 572-588. https://doi.org/https://doi.org/10.1509/jmr.15.0248 Park, S. (2020). Multifaceted trust in tourism service robots. Annals of Tourism Research , 81 , 102888. https://doi.org/https://doi.org/10.1016/j.annals.2020.102888 Pennebaker, J. W., Mehl, M. R., & Niederhoffer, K. G. (2003). Psychological aspects of natural language use: Our words, our selves. Annual review of psychology , 54 (1), 547-577. https://doi.org/https://doi.org/10.1146/annurev.psych.54.101601.145041 Pollitt, J. (1982). Moodiness: a heavenly problem? Observations on the pathology of moodiness. Journal of the Royal Society of Medicine , 75 (1), 7-16. https://doi.org/https://doi.org/10.1177/014107688207500104 Qian, K., & Yamada, Y. (2020). Exploring the role of the behavioral immune system in acceptability of entomophagy using semantic associations and food-related attitudes. Frontiers in Nutrition , 7 , 66. https://doi.org/https://doi.org/10.3389/fnut.2020.00066 Reich, T., Kaju, A., & Maglio, S. J. (2023). How to overcome algorithm aversion: Learning from mistakes. Journal of Consumer Psychology , 33 (2), 285-302. https://doi.org/ https://doi.org/10.1002/jcpy.1313 Richardson, D., Kinnear, B., Hauer, K. E., Turner, T. L., Warm, E. J., Hall, A. K., Ross, S., Thoma, B., Van Melle, E., & Collaborators, I. (2021). Growth mindset in competency-based medical education. Medical Teacher , 43 (7), 751-757. https://doi.org/https://doi.org/10.1080/0142159X.2021.1928036 Roesler, E., Heuring, M., & Onnasch, L. (2023). (hu) man-like robots: The impact of anthropomorphism and language on perceived robot gender. International Journal of Social Robotics , 15 (11), 1829-1840. https://doi.org/https://doi.org/10.1007/s12369-023-00975-5 Ryoo, Y., Jeon, Y. A., & Kim, W. (2024). The blame shift: Robot service failures hold service firms more accountable. Journal of Business Research , 171 , 114360. https://doi.org/https://doi.org/10.1016/j.jbusres.2023.114360 Schraft, R., Degenhart, E., Hagele, M., & Kahmeyer, M. (1993). New robot applications in production and service. Proceedings of 1993 IEEE/Tsukuba International Workshop on Advanced Robotics, Seo, H., Park, J., Bennis, M., & Debbah, M. (2023). Semantics-native communication via contextual reasoning. IEEE Transactions on Cognitive Communications and Networking . https://doi.org/10.1109/TCCN.2023.3250206 Shanks, I., Scott, M. L., Mende, M., van Doorn, J., & Grewal, D. (2024). Cobotic service teams and power dynamics: Understanding and mitigating unintended consequences of human-robot collaboration in healthcare services. Journal of the Academy of Marketing Science , 1-27. https://doi.org/https://doi.org/10.1007/s11747-024-01004-1 Su, L., Sengupta, J., Li, Y., & Chen, F. (2024). “Want” versus “Need”: How Linguistic Framing Influences Responses to Crowdfunding Appeals. Journal of Consumer Research , 50 (5), 923-944. https://doi.org/https://doi.org/10.1093/jcr/ucad033 Taylor, R. H. (2006). A perspective on medical robotics. Proceedings of the IEEE , 94 (9), 1652-1664. https://doi.org/https://doi.org/10.1109/JPROC.2006.880669 Torrent-Sellens, J., Jiménez-Zarco, A. I., & Saigí-Rubió, F. (2021). Do people trust in robot-assisted surgery? Evidence from Europe. International Journal of Environmental Research and Public Health , 18 (23), 12519. https://doi.org/https://doi.org/10.3390/ijerph182312519 Tussyadiah, I. P., Zach, F. J., & Wang, J. (2020). Do travelers trust intelligent service robots? Annals of Tourism Research , 81 , 102886. https://doi.org/https://doi.org/10.1016/j.annals.2020.102886 Wang, P. X., Wang, Y., & Jiang, Y. (2023). Gift or donation? Increase the effectiveness of charitable solicitation through framing charitable giving as a gift. Journal of Marketing , 87 (1), 133-147. https://doi.org/https://doi.org/10.1177/00222429221081506 Wang, X., Hwang, Y., & Guchait, P. (2023). When robot (vs. human) employees say “sorry” following service failure. International Journal of Hospitality & Tourism Administration , 24 (4), 540-562. https://doi.org/https://doi.org/10.1080/15256480.2021.2017812 Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of personality and social psychology , 54 (6), 1063. https://doi.org/https://doi.org/10.1037/0022-3514.54.6.1063 Zebrowski, R. L. (2020). Fear of a bot planet: Anthropomorphism, humanoid embodiment, and machine consciousness. Journal of Artificial Intelligence and Consciousness , 7 (01), 119-132. https://doi.org/https://doi.org/10.1142/S2705078520500071 Zhang, M., Cui, J., & Zhong, J. (2023). How consumers react differently toward humanoid vs. nonhumanoid robots after service failures: a moderated chain mediation model. International Journal of Emerging Markets . https://doi.org/https://doi.org/10.1108/IJOEM-06-2022-1023 Additional Declarations No competing interests reported. Supplementary Files Appendix.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6930952","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":513284081,"identity":"41158ca7-bcf4-4b55-87db-dfa1fa9e0564","order_by":0,"name":"Shanping Hu","email":"","orcid":"","institution":"Huainan Normal College","correspondingAuthor":false,"prefix":"","firstName":"Shanping","middleName":"","lastName":"Hu","suffix":""},{"id":513284082,"identity":"3898490d-56f9-4421-bc1a-333c1d839bf9","order_by":1,"name":"Dajun Yang","email":"","orcid":"","institution":"North Sichuan Medical College","correspondingAuthor":false,"prefix":"","firstName":"Dajun","middleName":"","lastName":"Yang","suffix":""},{"id":513284083,"identity":"272300e8-86a4-4e55-a1e3-85b343708eb4","order_by":2,"name":"Gui Gui","email":"","orcid":"","institution":"North Sichuan Medical College","correspondingAuthor":false,"prefix":"","firstName":"Gui","middleName":"","lastName":"Gui","suffix":""},{"id":513284084,"identity":"8e2d7f9c-7e2b-48c8-b4d9-90fa330b9513","order_by":3,"name":"Yunyun Wei","email":"","orcid":"","institution":"Huainan Normal College","correspondingAuthor":false,"prefix":"","firstName":"Yunyun","middleName":"","lastName":"Wei","suffix":""},{"id":513284085,"identity":"dd87eff6-5330-421f-868f-3521fc462cf7","order_by":4,"name":"Jun Cao","email":"","orcid":"","institution":"Huainan Normal College","correspondingAuthor":false,"prefix":"","firstName":"Jun","middleName":"","lastName":"Cao","suffix":""},{"id":513284086,"identity":"3ad8c111-5446-43ca-b8ea-a641842e6e6d","order_by":5,"name":"Daibo Xiao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0UlEQVRIie3QMQrCMBTG8QdCcYh2kxTFXiHSxUHwKg2Cs6OjRcjUA1i8hEd45UGmuAs6CK4d6tZFtK4ibdwc8pvff/gegOP8oxMAAsyYD4RYVvbJchwkWuZZapnUKBJkIup6FkX/nE6wWnfkRpuSgEHoD7A5CS46zlPjySQ9Hmg1hUm2j5sTcVog9hSTW14nOwaxOLcmcpM/FJcqLK7EPKtkgdRTImJgwC55b6GRiccctKifzNu31B+L7sX6yeZIt7KsZqE/bEk+8d/OHcdxnO9eQBBSsrjW+6kAAAAASUVORK5CYII=","orcid":"","institution":"City University of Macau","correspondingAuthor":true,"prefix":"","firstName":"Daibo","middleName":"","lastName":"Xiao","suffix":""}],"badges":[],"createdAt":"2025-06-19 11:53:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6930952/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6930952/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":99790492,"identity":"40ead7cb-9b91-403e-8923-7436ea33e29b","added_by":"auto","created_at":"2026-01-08 12:58:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":891943,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6930952/v1/7b734a44-b3d4-46b2-8321-392c63b73ebc.pdf"},{"id":91433692,"identity":"db971d47-32ad-4b60-8a5d-ec7e1b8dc42a","added_by":"auto","created_at":"2025-09-16 12:39:06","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":117563,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-6930952/v1/be772568aedd02f5793dd4cb.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"How Healthcare Chatbot Service Failures Affect Patient Trust?","fulltext":[{"header":"1 INTRODUCTION","content":"\u003cp\u003eThe healthcare chatbot is a multifunctional intelligent digital healthcare device (Kakhi et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). In addition to providing functions such as professional surgery(Han et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), postoperative rehabilitation guidance, and other auxiliary services(Guo et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Han et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), healthcare chatbots also play an important role in the patient care process (Guo \u0026amp; Li, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). There has been a great deal of resources invested in researching how to enhance the service of healthcare chatbots to help healthcare personnel reduce working pressure(Alhaffar \u0026amp; Janos, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) and meet the needs of patients in accessing healthcare. In recent years, the pressure on healthcare organizations and the workload of medical workers have increased dramatically, making it particularly important to actively utilize the service of healthcare chatbots (Guo et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). It has been told that healthcare chatbot service failure can have a huge impact on patients' psychology and attitudes(Choi et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, it is also reasonable to doubt whether the different types of healthcare chatbot service failure (process failure vs. outcome failure) actually affect patient trust in the healthcare chatbot. Nevertheless, academics in the field of robotic healing care have presented limited evidence to respond to this question.\u003c/p\u003e\u003cp\u003eRobotic service failure is a situation where the robot has problems performing a task or providing a service or is unable to complete the task. An extensive literature exists on the topic of healthcare chatbot services, and it has become a trend in research to promote the widespread healthcare chatbots (Gibelli et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Lee et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Although previous studies have highlighted that patient trust could influenced by the therapeutic effectiveness and nursing of healthcare chatbots (Liao et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Liu et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). For example, in a study by Torrent-Sellens et al. (\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), it was found that patient trust in healthcare chatbot therapy overrides rational decision-making. However, research on how healthcare chatbot service failure affects patient trust is somewhat flawed and insufficient. A recent study suggested that the healthcare chatbot can improve the way of patients trust in healthcare (Kerasidou, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Although previous studies have focused on the impact of healthcare chatbot service on patient trust (Asan et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Holth\u0026ouml;wer \u0026amp; van Doorn, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), not enough attention has been paid to whether healthcare chatbot service failure affects patient trust (Leo \u0026amp; Huh, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), so its value in terms of theory and nursing care has not been adequately recognized. This neglect is particularly unfortunate because healthcare chatbot service failure may trigger patients' skepticism towards healthcare organizations and question the overall competence of the chatbot, thus causing patients' disappointment, frustration and anxiety, which could ultimately affect therapeutic outcomes and the recovery process (Khaksar et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Shanks et al., \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). More importantly, in the field of healthcare chatbot service failure, the mechanism and effect of the type of healthcare chatbot service failure on patient trust have not been demonstrated clearly. Therefore, we need to answer the following questions: How does healthcare chatbot service failure affect patient trust? What are the internal mechanisms by which healthcare chatbot service failure affects patient trust?\u003c/p\u003e\u003cp\u003eTo fill the gap of previous research, this study developed a theoretical framework to investigate how to restore patient trust in healthcare chatbot service failure scenarios. Guided by the semantic framing, we argued that healthcare chatbot service failure reduces patient trust by affecting their mindset and altering their views and perceptions of the healthcare chatbot service. Despite previous research on the ability of healthcare chatbot service to influence patients' perceptions (Koutentakis et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), insufficient attention has been paid to the impact of healthcare chatbot service failure on patient trust. Based on this, we first proposed the impact of healthcare chatbot service failure on patient trust; second, we verified the interaction effect between patient mindset and healthcare chatbot service failure on patient trust; Last we explored the interaction effect between the linguistic expression type of the healthcare chatbot and healthcare chatbot service failure on patient trust.\u003c/p\u003e\u003cp\u003eThis study makes the following important contributions to service robots and patient trust. Although it has been common for studies to address healthcare chatbot service (Choi et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), few studies have focused on the types of healthcare chatbot service failure and their impact on patient trust. Therefore, this study is an early attempt to link the healthcare chatbot service failure with the patient trust, which further clarified the internal mechanisms and boundary conditions between the two, enriching the existing literature on healthcare chatbot services and patient trust. Second, this study explored the interaction effect between patient mindset (Cowger et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and healthcare chatbot service failure. Patients' growth mindset and fixed mindset towards the healthcare chatbot service may have differential effects on patient trust, and this could also better explain the impact of changed patient mindset on patient trust with healthcare chatbot service failure. Last, this study examined the interaction effect between the linguistic expression type of the healthcare chatbot service and healthcare chatbot service failure. Robot language is mostly set by programming algorithms, but the linguistic expression type significantly affects patient trust (Foster, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), and an in-depth exploration could better illustrate the internal mechanisms between healthcare chatbot service failure and patient trust. In summary, the theoretical contribution of this study is to provide a systematic explanation and understanding of how healthcare chatbot service failure combines with patient mindset and linguistic expression type, which in turn affects patient trust. In fact, this study also provides valuable guidance for the development of healthcare chatbot service models, the maintenance of doctor-patient relationships, and the service capabilities of hospitals.\u003c/p\u003e"},{"header":"2 THEORETICAL FRAMEWORK AND HYPOTHESIS DERIVATION","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 AI AND SERVICE ROBOTS\u003c/h2\u003e\u003cp\u003eWith the continuous development of science and technology, robots are increasingly used in various fields. Especially in the service industry, robots have become an important tool for providing efficient and convenient services (Brooks, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e1991\u003c/span\u003e; Luo et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). At the same time, the drawbacks of robot services have gradually emerged and attracted extensive attention from academics (Moriuchi \u0026amp; Murdy, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Ryoo et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Based on the related literature, first we categorized the current literature on the connotation of AI services, and then provided an overview of the existing research in terms of both positive and negative factors that affect consumers' adoption of AI in the service industry.\u003c/p\u003e\u003cdiv id=\"Sec4\" class=\"Section3\"\u003e\u003ch2\u003e2.1.1 AI SERVICE\u003c/h2\u003e\u003cp\u003eA large amount of literature has discussed the connotation of AI services. The definition of AI services is rooted in AI technology. According to Liu et al. (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), AI refers to intelligent programs, algorithms, systems, and machines, which can present different characteristics of human intelligence (Holzinger et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Huang et al. (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) specified 4 types of intelligence of AI, including mechanical, analytical, intuitive and empathetic. Based on this, Doupe et al. (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) defined robot services as system-based autonomy and adaptability that can interact, communicate and deliver services to organized customers. Researchers also classified service robots according to the dimensions of functionality, socio-emotion, and relationship. Recently, Bock et al. (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) tend to identify AI services as a technological configuration that deliver service value through perception, learning, decision-making, and action realization with flexible adaptation.\u003c/p\u003e\u003cp\u003eIntegrating the existing literature, this study argued that AI services can be defined as follows. It is a process by which algorithm- and technology-driven intelligent devices (which may have anthropomorphic appearances, languages and personalities) provide service value to customers with communication and interaction. With the continuous improvement and maturation of AI technology, future AI services will be presented with trend of intelligence, personalization and emotionalization.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section3\"\u003e\u003ch2\u003e2.1.2 SERVICE ROBOT\u003c/h2\u003e\u003cp\u003eService robots is a relatively new field whose definition and scope are evolving over time(Chi et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Taking into account different definitions, service robots are described as robots that are able to perform useful tasks. These performed tasks can either serve humans or devices (Becker et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The development of service robots has reduced the workload of healthcare workers and helped them carry out complex tasks. The earliest definition of a service robot was proposed by Schraft in 1993. They defined it as \u0026ldquo;A service robot is a freely programmable kinematic device that performs services semi-or fully automatically. Services are tasks that do not contribute to the industrial manufacturing of goods but are the execution of useful work for humans and equipment\u0026rdquo; (Schraft et al., \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e1993\u003c/span\u003e). As service robots have evolved, researchers have come up with additional definitions. The International Organization for Standardization defines service robots as \u0026ldquo;a robot that performs useful tasks for humans or equipment, excluding industrial automation applications\u0026rdquo;. The International Federation of Robotics, on the other hand, emphasizes the autonomy of robots, defining service robots as \u0026ldquo;a service robot is a robot which operates semi- or fully autonomously to perform services useful to the well-being of humans and equipment, excluding manufacturing operations\u0026rdquo;(Holland et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, as the term \"Service Robot\" continues to evolve, the its definition has become ambiguous, because there is a crossover between the industrial and service sectors (Dupont et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTo sum up, this study argued that a service robot is capable of performing useful tasks that is intended to provide services to humans and devices. The definition and scope of service robots have evolved over time, but their core goal has always been to deliver helpful services to humans and devices.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section3\"\u003e\u003ch2\u003e2.1.3 FACTORS AFFECTING THE APPLICATION OF SERVICE ROBOT\u003c/h2\u003e\u003cp\u003eExisting literature focused on how AI services affect customer psychology and thus act on customer behavior from the perspective of human-machine interaction. Part of those literatures explored the positive incentives that motivate consumers to adopt AI from different aspects, including consumer trust in anthropomorphic robots, and their perception of anthropomorphic robots' enthusiasm.\u003c/p\u003e\u003cp\u003eSpecifically, Chuah et al. (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) demonstrated experimentally that having more anthropomorphic features in consumer robots would increase customers' perception of enthusiasm. Trust in intelligent algorithms was higher when humans performed those less threatened tasks. An experimental study by Barone et al. (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) demonstrated that anthropomorphism has a function in guiding consumers to make positive evaluations of robots when they have failed to provide service. Because customers' anthropomorphic perceptions of service robots facilitated their intention to use the service robots. It was found that anthropomorphic robots can repair customers' enthusiasm from service remediation through apologies and explanations, which in turn leads to renewed satisfaction, and consumer trust enhances their willingness to use humanoid service robots (Hwang et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Although existing studies have found an \"algorithm aversion\" (Reich et al., \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), some academic researchers have suggested that the positive impact of anthropomorphism on transaction switching has been progressively explored, as consumers' increasing acceptance of the algorithms in their daily lives (Roesler et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThere were also studies discussed the negative factors that affect customers' adoption of AI, such as the lack of compassion and empathy of AI and consumers' lack of control over AI services(Dang \u0026amp; Liu, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The discomfort may be caused by overly anthropomorphic robots and the consumers' personality could be neglected by AI(Zebrowski, \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Meanwhile, the literature has found that compassion and empathy are two human qualities that are lacking in machines. Therefore, customers' distrust, feelings of being ignored and unappreciated, and closed mind that resist change also limit the adoption of AI services (Kuchenbrandt et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). In an earlier study, Mende et al. (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) found that consumers of anthropomorphic service robots, even when faced with flawless AI and robotic services, may be affected by speciesism and consequently avoid the use of service robots. In addition, it has also been found that service robots have more challenges than humans in interpreting consumers' unique characteristics and environments in providing healthcare services, which has led to consumers' resistance to healthcare AI services (Huang et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.2 ROBOT SERVICE FAILURE\u003c/h2\u003e\u003cp\u003eRobot service failure can be defined as any form of actual or perceived misfortune, error, or problem that occurs during the stage of consumer experience (Arikan et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This failure results in the behavior or service performed by the robotic system deviating from the desired, normal or correct function (Zhang et al., \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Robot service failures include both perceivable failures and actual failures caused by correct behaviors performed by the robot in accordance with its programming (Hu et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Liu et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Robot service failures are inevitable, because various weaknesses in its hardware, structure, and applications can lead to service failures. These weaknesses may include technical limitations, design flaws, and software bugs of the robot. These issues can trigger consumer dissatisfaction and negative word-of-mouth(Leo \u0026amp; Huh, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn order to specifically classify robot service failures, existing research has mainly divided it into two dimensions, process failure and outcome failure in interpersonal service contexts(Choi et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Process failure refers to deficiencies in service delivery, such as untimely or incorrect robot responses. Outcome failure, on the other hand, refers to the failure of the robot to fulfil the basic service content, such as delivering the wrong goods. By dividing robot service failure into process failure and outcome failure, we can better understand the problems in robot service and take appropriate measures to improve service quality. For process failure, we can improve the service efficiency by upgrading robots' response speed and accuracy. For outcome failures, we can enhance the basic service content of the robot to ensure that consumers receive the services they expect(P. X. Wang et al., \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; X. Wang et al., \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). By categorizing different types of failures and proposing specific solutions, we can improve the quality of robot services and user experience.\u003c/p\u003e\u003cp\u003eHowever, robot service failure may have more serious consequences in the healthcare domain as it relates to the health and lives of patients. Therefore, discussion on healthcare chatbot service failure is particularly important. Unfortunately, previous studies have not deeply discussed the antecedents of healthcare chatbot service failure (process failure vs. outcome failure), which has led to the absence of a clear answer in the academic community, on how to regain patient trust in a context of healthcare chatbot service failure.\u003c/p\u003e\u003cp\u003eThis study inferred that process failures can usually be corrected, whereas outcome failures are probably unchangeable. When patients realize that process failures can be fixed, patients will have more confidence and patience to wait for it to be resolved, rather than feeling despair and disappointment due to outcome failures. In addition, the process failure can be handled in a way that the doctor and robot are open to the patients with transparency and honesty, about the cause of the problem and the solution. It can increase the patient trust in the doctor or robot and avoid disappointment.\u003c/p\u003e\u003cp\u003eBased on the above analyses, the following hypothesis was proposed.\u003c/p\u003e\u003cp\u003e\u003cb\u003eH1: Healthcare chatbot process failure (vs. outcome failure) has less impact on patient trust.\u003c/b\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e2.3 INTERACTION EFFECT OF PATIENT MINDSET\u003c/h2\u003e\u003cp\u003ePatient mindset and robot service failure are two factors that are closely related to patient trust. Patient mindset refers to the patient's mental state and attitude in receiving healthcare services, while robot service failure refers to the problems or mistakes made by the robot in providing healthcare services (Chiu et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e1997\u003c/span\u003e). The interaction between these two factors has a significant impact on the emergence and maintenance of patient trust. The interaction between patient mindset and robot service failure on patient trust is a complex process with multiple levels and dimensions (Meng et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Patient mindset can be broadly categorized into growth mindset and fixed mindset (Chiu et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e1997\u003c/span\u003e), while robot service failure includes both process failure and outcome failure. The interaction between these two plays a crucial role in the emergency and maintenance of patient trust.\u003c/p\u003e\u003cp\u003eFirst, we explored the interaction between growth patient mindset and robot service failure. A growth patient mindset implies that patients tend to view the healthcare process positively, believing that through continuous learning and effort, they can overcome the difficulties associated with their illness (Fissell et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Patients with this mindset are more likely to view process failures of robot services as opportunities to learn and improve rather than as mere failures when faced with them (Grant, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). They will understand and accept the limitations in its initial stages, while believing these issues will be resolved as the technology continues to advance (Ching et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). As a result, patients with growth mindset are more likely to maintain a high level of trust even if there are process failures in robot services.\u003c/p\u003e\u003cp\u003eHowever, even patients with growth mindset may face trust challenges when there are robot services outcome failures, as this is directly related to their treatment outcome and health status, which is their crucial concern (Joseph et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). It is also difficult for patients who hold a growth mindset to accept failed treatments or deterioration of their condition due to errors in robot services(Chao et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In this case, patients may become skeptical about the reliability and effectiveness of the robots, thus affecting their trust.\u003c/p\u003e\u003cp\u003eNext, we analyzed the interaction between fixed patient mindset and robot service failure. Fixed patient mindset manifests itself as patients holding more fixed views and expectations about their health conditions and medical processes. Such patients may be more prone to negative emotions and distrust when faced with process failures of robot services. They may view process failure as an inherent flaw rather than a normal part of the technological development process(Richardson et al., \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Thus, for patients with fixed mindset, their trust in robot services may be significantly reduced by process failure.\u003c/p\u003e\u003cp\u003eAs for outcome failure, the impact may be even more severe for patients with fixed mindset. Such patients usually have high expectations for medical outcomes and may feel extremely disappointed and angry when the robot service fails to achieve the desired treatment outcome (Desveaux \u0026amp; Ivers, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). They may blame the outcome failure on the deficiencies of the robotic technology and thus lose trust in the robot service altogether.\u003c/p\u003e\u003cp\u003eBased on the above analyses, the following hypothesis was proposed.\u003c/p\u003e\u003cp\u003e\u003cb\u003eH2: Healthcare chatbot service failure (process failure vs. outcome failure) and patient mindset (growth vs. fixed) have an interaction effect on patient trust.\u003c/b\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e2.4 SEMANTIC FRAME OF \u0026ldquo;WANT\u0026rdquo; VS. \u0026ldquo;NEED\u0026rdquo;\u003c/h2\u003e\u003cp\u003eThe semantic frame is a structure used to represent the relationships between words and sentences in language, which describes the semantic relationships between lexical items, including hypernym and hyponym, parts and wholes, attributes and entities, and so on. Semantic frames can help us understand and interpret meaning in our language.\u003c/p\u003e\u003cp\u003eIn recent years, a large number of academic researchers have focused on the impact of semantic frames on consumer behavior and have demonstrated its prominent importance in marketing and advertising. By choosing the right words or phrases, marketers can convey different emotions and intentions to influence consumers' purchasing decisions and attitudes. For example, when recommending a product or service, using \"I recommend it\" is more persuasive than \"I like it\" (Packard \u0026amp; Berger, \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The former implies a higher level of professionalism and reliability, whereas the latter only expresses personal preferences. Similarly, when asking for help or support, the use of \"will you\" is more likely to receive a positive response than \"can you\" (Jia et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The former implies the other person's willingness and ability, whereas the latter only expresses the other person's ability. These small semantic changes may have far-reaching effects on consumers. Therefore, marketers need to choose and use words carefully to ensure that they convey the emotions and intentions they want.\u003c/p\u003e\u003cp\u003eFrom the findings of psycholinguistic research, it is clear that language can effectively influence people's perceptions of certain people or things. Research has shown that frequent use of the first-person singular has been shown to be neurotic (Pennebaker et al., \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). The research on event-related potentials provides direct neuroscientific evidence that semantics affects individuals' behavior. In recent years, academic researchers have found that semantic symbols can affect individuals' memory, which in turn influences how they act (Seo et al., \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). As a result, semantic associations have gained the attention of marketing scholars. Qian and Yamada (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) found that semantic associations have a direct correlation with consumers' attitudes towards products. There are indications that semantic frames can effectively influence individual behavior.\u003c/p\u003e\u003cp\u003eBased on findings from the literature on semantic frame, we questioned to what extent a healthcare chatbot dialogue request framed as \"need\" vs. \"want\" affects patient trust\u0026ndash;\"need\" vs. \"want\" triggers different patients' perceptions regarding the professionalism of the healthcare chatbot (Su et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). We further argued that due to these different perceptions, the semantic frames of \"need\" and \"want\" have different mechanisms for influencing patient trust.\u003c/p\u003e\u003cp\u003eIn research on consumption, academics generally supported the idea that \"need\" refer to the things that consumers must have in order to maintain their basic needs in life. These \"need\" are the items that people must purchase in order to satisfy their basic survival and living needs, such as food, clothing, shelter, and so on. It can be seen that needs exist to satisfy the basic requirements of life, without which people's lives would be difficult or even impossible to maintain. \"Want\", on the other hand, refers to consumers' desire or longing for certain items or experiences. These are non-essential items or experiences that do not exist to fulfil basic life needs. The wanted things are usually like luxury items, recreational activities and travelling. They are not essential to life, but can provide additional enjoyment and satisfaction (Su et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Therefore, the difference between \"need\" and \"want\" lies in the fact that \"need\" is something that must be possessed in order to satisfy the basic necessities of life, whereas \"want\" is a desire for a non-essential object or experience. \"want\" is a desire for non-essential objects or experience. A \"need\" is basic, while a \"want\" is additional. In consumption decision-making, rational consumers will first satisfy their needs before considering meeting their wants (Baker, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThis stratification suggests that the \"need\" and \"want\" sent to the patient, when the healthcare chatbot service failure occurs, may affect patients' willingness for further interaction. If the healthcare chatbot is able to satisfy the patients' basic needs (\"need\"), the patient may be more willing to continue interacting with the robot because they perceive their ability to provide the necessary help and support. On the contrary, if the healthcare chatbot only satisfies the patients' additional needs (\"wants\"), patients may perceive their services as not necessary, thus diminishing patients' willingness for further interaction.\u003c/p\u003e\u003cp\u003eOn the basis of these considerations, we proposed the central premise of this study: the use of \"need\" (or \"want\") in patient interactions with a healthcare chatbot will affect the extent to which the patient trust in the healthcare chatbot is perceived differently. As assumed in the previous discussion, \"need\" and \"want\" requests do not imply the same constraints. The \"want\" semantic frame implies greater adequacy and higher capacity for action without constraints. Therefore, we argued that the use of \"need\" vs. \"want\" in healthcare chatbot requests will affect patients' different inferences about the trustworthiness of the healthcare chatbot.\u003c/p\u003e\u003cp\u003e(To our knowledge) this inference has never been directly confirmed. However, there is some indicative evidence for healthcare chatbot-patient interactions. For example, Su et al. (\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) found that \"need\" vs. \"want\" affects differences in customer perceptions, which can lead to different individual willingness for donation.\u003c/p\u003e\u003cp\u003eThe current research tested the difference in patient trust in a healthcare chatbot service failure scenario, after sending a sentence containing \"need\" vs. \"want\" to a patient. We argued that it affects patient trust in the healthcare chatbot because the semantic frames of \"need\" and \"want\" imply different semantic signals. It is important to reiterate that, consistent with past research, \"need\" and \"want\" in the current study are not the result of a binary classification, but rather a continuum. Thus, \"need\" does not imply an absolute service relationship between the healthcare chatbot and the patient, whereas \"want\" implies a more conciliatory tone for the healthcare chatbot. Argument in this study is that \"want\" positively affects patient trust more than \"need\".\u003c/p\u003e\u003cp\u003eBased on the above analyses, the following hypothesis was proposed.\u003c/p\u003e\u003cp\u003e\u003cb\u003eH3: Healthcare chatbot service failure (process failure vs. outcome failure) and linguistic expression type (need vs. want) have an interaction effect on patient trust.\u003c/b\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"3 OVERVIEWS OF THE STUDY","content":"\u003cp\u003eTo test the above three hypotheses, this study verified the impact of healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, along with its internal mechanisms and boundary conditions, by conducting three related experiments. Experiment 1 verified the main effect of healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, verifying H1; Experiment 2 attempted to investigate the interaction effect between patient mindset (growth vs. fixed) and healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, verifying H2; and Experiment 3 analyzed the interaction effect between linguistic expression types (need vs. want) and healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, verifying H3. In all experiments, we firstly asked all the participants to read the definition of the healthcare chatbot to enrich their knowledge of it. Then, to better manipulate the healthcare chatbot service failure, we used different situation statements as stimulus material for each experiment.\u003c/p\u003e"},{"header":"4 EXPERIMENT 1: MAIN EFFECT OF ROBOT SERVICE-PATIENT TRUST","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e4.1 EXPERIMENTAL DESIGN\u003c/h2\u003e\u003cp\u003eExperiment 1 aimed to explore the main effect of robot services on patient trust. We designed a one-way between-subjects design (robot service: process failure vs. outcome failure), and recruited 300 participants on credamo (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.credamo.com/\u003c/span\u003e\u003cspan address=\"https://www.credamo.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), of which 147 (49.0%) were male and 153 (51.0%) were female. The age distribution of all the participants was 24.0% aged 18\u0026ndash;25 years, 30.0% aged 26\u0026ndash;40 years, 24.3% aged 41\u0026ndash;60 years and 21.7% aged 61 years and above. All participants were randomly divided into two groups and assigned to two different experimental scenarios (outcome failure group Vs. process failure group). The process failure group with a total of 150 participants got the corresponding answers although questions arose during the process; the outcome failure group with a total of 150 participants failed to get the corresponding answers because the questions arose during the process.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e4.2 EXPERIMENTAL PROCESS\u003c/h2\u003e\u003cp\u003eFirst, all participants were guided to imagine that they were in a medical institution, and after typing in a medical question to the healthcare chatbot, they took a long time and repeatedly pressed the \"Answer Output\" button on the screen. Participants in the process failure group got the answer to the question after several attempts. However, the participant in the outcome failure group got the hospital's information after several attempts. Next, we asked the participants, \"Do you agree that you were satisfied with your interaction with the healthcare chatbot?\", \"Do you agree that you developed a good feeling about the healthcare chatbot? \", \"Do you agree that you felt a strong sense of positive engagement?\", and \"Do you agree that in interacting with the healthcare chatbot, it played an influential interactive role for the patient?\" (1\u0026thinsp;=\u0026thinsp;Strongly Disagree, 7\u0026thinsp;=\u0026thinsp;Strongly Agree) to measure participants' trust. The scale was adapted from the Trust Scale developed by Jin and Eastin (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eBased on the study of Watson et al. (\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e1988\u003c/span\u003e), we found that the emotional state of the patient has a significant impact on the patient trust, and in order to enhance the accuracy of the experimental results, we need to manipulate the emotions of the participants. We asked the participants \"Do you have a good emotion and are happy, pleased, and satisfied?\" (1\u0026thinsp;=\u0026thinsp;Strongly Disagree, 7\u0026thinsp;=\u0026thinsp;Strongly Agree), a question taken from the scale of Watson et al. (\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e1988\u003c/span\u003e). Last, we collected demographic information about the participants (Cronbach's α\u0026thinsp;=\u0026thinsp;0.655).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e4.3 EXPERIMENTAL RESULTS\u003c/h2\u003e\u003cp\u003eThere was a main effect test. We employed one-way ANOVA with healthcare chatbot service failure as the independent variable and patient trust as the dependent variable. The results showed that patient trust in the robot service process failure group (M\u0026thinsp;=\u0026thinsp;5.02, SD\u0026thinsp;=\u0026thinsp;1.002) was significantly higher than that in the outcome failure group (M\u0026thinsp;=\u0026thinsp;3.55, SD\u0026thinsp;=\u0026thinsp;0.843), F (1, 298)\u0026thinsp;=\u0026thinsp;189.754, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001. H1 was verified.\u003c/p\u003e\u003cp\u003eThere was a control variable analysis. Based on the study of Pollitt (\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e1982\u003c/span\u003e), we found that emotion is an important factor that could affect individual trust. Therefore, we conducted an analysis of covariance (ANCOVA) with patient trust as the dependent variable, healthcare chatbot service as the independent variable, and patients' emotions as the covariate. The results of the experiment showed that patient emotion did not have a significant effect on patient trust (F (1, 298)\u0026thinsp;=\u0026thinsp;28.308, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Therefore, patient emotion did not affect the results of the experiment, and H1 was again verified.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e4.4 DISCUSSION\u003c/h2\u003e\u003cp\u003eExperiment 1 verified that healthcare chatbot service has a significant positive impact on patient trust, and the process failure of healthcare chatbot service affects patient trust more than the outcome failure. We also verified that there was no significant effect of patient emotions on patient trust so as to enhance the reliability and accuracy of the experiment. Despite the above findings, Experiment 1 failed to further explore the internal mechanisms and boundary conditions between the healthcare chatbot and patient trust. Experiment 2 introduced patient mindset as a moderating variable to explore the interaction effect between patient mindset and robot service failure on patient trust.\u003c/p\u003e\u003c/div\u003e"},{"header":"5 EXPERIMENT 2: INTERACTION EFFECT BETWEEN PATIENT MINDSET AND HEALTHCARE CHATBOT SERVICE FAILURE","content":"\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003e5.1 EXPERIMENTAL DESIGN\u003c/h2\u003e\u003cp\u003eExperiment 2 aimed mainly at: first, re-validating the main effect; and second, exploring the interaction effect between patient mindset (growth vs. fixed) and healthcare chatbot service failure on patient trust. We designed a 2 (robot service: process failure vs. outcome failure) X 2 (patient mindset: growth vs. fixed) ANOVA and recruited 600 participants on credamo (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.credamo.com/\u003c/span\u003e\u003cspan address=\"https://www.credamo.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), of which 301 (50.2%) were male and 299 (49.8%) were female. The age distribution of all the participants was 24.5% aged 18\u0026ndash;25 years, 27.8% aged 26\u0026ndash;40 years, 25.5% aged 41\u0026ndash;60 years and 22.2% aged 60 years and above. All participants were then assigned to two different experimental scenarios (outcome failure group Vs. process failure group). The 302 participants in the process failure group got the corresponding answers although the questions arose during the process. In the process failure group, we used the scenario material to subdivide them into the process failure-growth group (N\u0026thinsp;=\u0026thinsp;153) and the process failure-fixed group (N\u0026thinsp;=\u0026thinsp;149) with the patient's mindset. The 298 participants in the outcome failure group failed to get the corresponding answers because the questions arose during the process. In the outcome failure group, we used the scenario material to subdivide them into the outcome failure-growth group (N\u0026thinsp;=\u0026thinsp;148) and the outcome failure-Fixed group (N\u0026thinsp;=\u0026thinsp;150).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003e5.2 EXPERIMENTAL PROCESS\u003c/h2\u003e\u003cp\u003eFirst, all participants were guided to imagine that they were in a medical institution, and after typing in a medical question to the healthcare chatbot, they took a long time and repeatedly pressed the \"Answer Output\" button on the screen. Participants in the process failure group got the answer to the question after several attempts. However, the participant in the outcome failure group got the hospital's information after several attempts. We all told the participants of each fixed group that the programming of the healthcare chatbot can be continuously learnt and improved, but all robots have limitations in their growth, and its service capability will be fixed for and will not grow a long time to come; while we told the participants of each growth group that the programming of the robot can be continuously learnt and improved, and the service and communication methods are gradually humanized and scientized, and are able to satisfy personal companionship (\u0026ldquo;need\u0026rdquo;). Second, we asked the participants in the fixed group, \"Do you agree that no matter how capable the healthcare chatbot is, we can always change it more and help it grow?\", and \"Do you agree that we can always change the healthcare chatbot's capabilities significantly?\" (1\u0026thinsp;=\u0026thinsp;Strongly Disagree, 7\u0026thinsp;=\u0026thinsp;Strongly Agree); and asked participants in the growth group, \"Do you agree that the healthcare chatbot has some limits to its capabilities, that you can't ask too much of it, and that its evolution is slow?\", and \"Do you agree that a healthcare chatbot can learn new things, but we can't really change the basic level of competence of a healthcare chatbot?\" (1\u0026thinsp;=\u0026thinsp;strongly disagree, 7\u0026thinsp;=\u0026thinsp;strongly agree) Patient mindset was measured using two different sets of questions adapted from the scale of Chiu et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e1997\u003c/span\u003e). Next, we asked the participants, \"Do you agree that you were satisfied with your interaction with the healthcare chatbot?\", \"Do you agree that you developed a good feeling about the healthcare chatbot?\", \"Do you agree that you felt a strong sense of positive engagement?\", and \"Do you agree that in interacting with the healthcare chatbot, it played an influential interactive role for the patient?\" (1\u0026thinsp;=\u0026thinsp;Strongly Disagree, 7\u0026thinsp;=\u0026thinsp;Strongly Agree) to measure participants' trust. The scale was adapted from the Trust Scale developed by Jin and Eastin (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Last, we collected demographic information about the participants (Cronbach's α\u0026thinsp;=\u0026thinsp;0.68).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003e5.3 EXPERIMENTAL RESULTS\u003c/h2\u003e\u003cp\u003eThere was a main effect test. We conducted a one-way ANOVA with healthcare chatbot service failure as the independent variable and patient trust as the dependent variable. The experimental results showed that patient trust in process failure group (M\u0026thinsp;=\u0026thinsp;4.36, SD\u0026thinsp;=\u0026thinsp;1.062) was significantly higher than that in outcome failure group (M\u0026thinsp;=\u0026thinsp;4.1527, SD\u0026thinsp;=\u0026thinsp;1.272), F (1, 598)\u0026thinsp;=\u0026thinsp;4.702, p\u0026thinsp;=\u0026thinsp;0.031. H1 was verified.\u003c/p\u003e\u003cp\u003eThere was a manipulation check. We conducted a one-way ANOVA with patient mindset as the independent variable and patient trust as the dependent variable. The results of the experiment showed that M \u003csub\u003egrowth group\u003c/sub\u003e = 4.931, SD \u003csub\u003egrowth group\u003c/sub\u003e = 1.018; M \u003csub\u003efixed group\u003c/sub\u003e = 3.578, SD \u003csub\u003efixed group\u003c/sub\u003e = 0.9; and F (1, 598)\u0026thinsp;=\u0026thinsp;296.815, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001. It can be seen that the patient trust in the growth mindset group was significantly higher than that in the fixed mindset group. Therefore, the manipulation in Experiment 2 was successful.\u003c/p\u003e\u003cp\u003eThere was an interaction effect test. We conducted a 2 (robot service: process failure vs. outcome failure) X 2 (patient mindset: growth vs. fixed) ANOVA with patient trust as the dependent variable, in an attempt to validate the interaction effect between patient mindset and healthcare chatbot service failure on patient trust. The results showed that healthcare chatbot service failure had a significant effect on patient trust (F (1, 598)\u0026thinsp;=\u0026thinsp;6.355, p\u0026thinsp;\u0026lt;\u0026thinsp;0.012), patient mindset had a significant effect on patient trust (F (1, 598)\u0026thinsp;=\u0026thinsp;306.081, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and the interaction between patient mindset and healthcare chatbot service failure had a significant effect on patient trust (F (1, 598)\u0026thinsp;=\u0026thinsp;14.458, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). It can be seen that the interaction between patient mindset and healthcare chatbot service failure had a significant effect on patient trust, which validated H2.(shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e[Insert Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e about here]\u003c/p\u003e\u003cp\u003eThere was a control variable analysis. We considered that gender may affect patient trust after healthcare chatbot service failure Barone et al. (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Therefore, we employed one-way ANOVA with gender as the independent variable and patient trust as the dependent variable. The results showed that patient gender had no significant effect on patient trust (F (1,598)\u0026thinsp;=\u0026thinsp;0.022, p\u0026thinsp;=\u0026thinsp;0.882). Therefore, patient gender did not affect the results of the experiment and H1 was again validated.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003e5.4 DISCUSSION\u003c/h2\u003e\u003cp\u003eExperiment 2 verified the interaction effect between patient mindset and healthcare chatbot service failure on patient trust. The experimental results showed that in the face of healthcare chatbot service failure, patient trust in the growth mindset group was significantly higher than that in the fixed mindset group. Specifically, with the healthcare chatbot service process failure, the difference between the patient trust in the growth mindset group and the fixed mindset group is smaller, showing a tendency to intersect; with the healthcare chatbot service outcome failure, the difference between the patient trust in the growth mindset group and the fixed group is larger, showing a tendency to diverge. At the same time, we excluded the impact of patients' gender on the experimental results, which enhanced the accuracy of the experiment. Despite the above findings in Experiment 2, the impact of the potential factors from the healthcare chatbot service itself on patient trust has not been further explored. Therefore, in order to address the above research gap, Experiment 3 introduced a healthcare chatbot linguistic expression type to investigate its interaction effect with healthcare chatbot service failure on patient trust.\u003c/p\u003e\u003c/div\u003e"},{"header":"6 EXPERIMENT 3: INTERACTION EFFECT BETWEEN LINGUISTIC EXPRESSION TYPE AND HEALTHCARE CHATBOT SERVICE FAILURE","content":"\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e\u003ch2\u003e6.1 EXPERIMENTAL DESIGN\u003c/h2\u003e\u003cp\u003eExperiment 3 aimed to investigate the interaction effect of linguistic expression type and healthcare chatbot service failure on patient trust. We designed 2(healthcare chatbot service failure: process failure vs. outcome failure) X 2(linguistic expression type: need vs. want) ANOVA and recruited 520 participants on credamo (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.credamo.com/\u003c/span\u003e\u003cspan address=\"https://www.credamo.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), of which 259 (49.8%) subjects were male and 261 (50.2%) subjects were female. The age distribution of all participants was 20% aged 18\u0026ndash;25 years, 31.5% aged 26\u0026ndash;40 years, 25.2% aged 41\u0026ndash;60 years and 23.3% aged 60 years and above. All participants were then assigned to two different experimental scenarios (outcome failure group Vs. process failure group). The participants in the process failure group got the corresponding answers although the questions arose during the process. In the process failure group, we used scenario materials of linguistic expression type to subdivide participants into process failure-need group (N\u0026thinsp;=\u0026thinsp;94) and process failure-want group (N\u0026thinsp;=\u0026thinsp;146). The participants in the outcome failure group failed to get the corresponding answers because the questions arose during the process. In the outcome failure group, we used the scenario materials of linguistic expression type to subdivide the participants into process failure-need group (N\u0026thinsp;=\u0026thinsp;150) and process failure-want group (N\u0026thinsp;=\u0026thinsp;130).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec23\" class=\"Section2\"\u003e\u003ch2\u003e6.2 EXPERIMENTAL PROCESS\u003c/h2\u003e\u003cp\u003eFirst, all participants were guided to imagine that they were in a medical institution, and after typing in a medical question to the healthcare chatbot, they took a long time and repeatedly pressed the \"Answer Output\" button on the screen. Participants in the process failure group got the answer to the question after several attempts. However, the participant in the outcome failure group got the hospital's information after several attempts. Then, in the need group, the healthcare chatbot asked the participant: \"I'm sorry, I made a mistake in my understanding, I need you to retype the question, can we proceed with the diagnosis?\"; and in the want group, the healthcare chatbot asked the participant: \"I'm sorry, I made a mistake in my understanding, I want to ask for your understanding, can you retype the question, and let's proceed with the diagnosis?\". Then, we asked the participants, \"Do you agree that you were satisfied with your interaction with the healthcare chatbot?\", \"Do you agree that you developed a good feeling about the healthcare chatbot?\", \"Do you agree that you felt a strong sense of positive engagement?\", and \"Do you agree that in interacting with the healthcare chatbot, it played an influential interactive role for the patient?\" (1\u0026thinsp;=\u0026thinsp;Strongly Disagree, 7\u0026thinsp;=\u0026thinsp;Strongly Agree) to measure participants' trust. The scale was adapted from the Trust Scale developed by Jin and Eastin (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eBased on the study of Kamide et al. (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), we found that psychological acceptance of the robot may have an impact on patient trust. Therefore, to manipulate patients' psychological acceptance, we asked participants \"Do you agree that the robot made you feel uncomfortable?\" and \"Do you agree that robots seem to reduce the value of the human healthcare workers?\" (1\u0026thinsp;=\u0026thinsp;Strongly Disagree, 7\u0026thinsp;=\u0026thinsp;Strongly Agree), which was adapted from the Psychological Acceptance Scale developed by Kamide et al. (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Last, we collected demographic information about the participants (Cronbach's α\u0026thinsp;=\u0026thinsp;0.634).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec24\" class=\"Section2\"\u003e\u003ch2\u003e6.3 EXPERIMENTAL RESULTS\u003c/h2\u003e\u003cp\u003eThere was a main effect test. We conducted a one-way ANOVA with healthcare chatbot service failure (process failure Vs. outcome failure) as the independent variable and patient trust as the dependent variable. The experimental results showed that patient trust in service process failure group (M\u0026thinsp;=\u0026thinsp;4.511, SD\u0026thinsp;=\u0026thinsp;1.041) was significantly higher than that in service outcome failure group (M\u0026thinsp;=\u0026thinsp;3.957, SD\u0026thinsp;=\u0026thinsp;1.291), F (1, 518)\u0026thinsp;=\u0026thinsp;28.375, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001. H1 was validated.\u003c/p\u003e\u003cp\u003eThere was a manipulation check. We conducted a one-way ANOVA with the linguistic expression types (need vs. want) as the independent variable and patient trust as the dependent variable. The results of the experiment showed that M \u003csub\u003eneed\u003c/sub\u003e=4.956, SD \u003csub\u003eneed\u003c/sub\u003e=0.902; M \u003csub\u003ewant\u003c/sub\u003e=3.371, SD \u003csub\u003ewant\u003c/sub\u003e=0.941; F (1, 518)\u0026thinsp;=\u0026thinsp;383.344, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001. It can be seen that patient trust in the want group was significantly higher than in the need group, and the manipulation in experiment 3 was successful.\u003c/p\u003e\u003cp\u003eThere was an interaction effect test. We analyzed the interaction effect of healthcare chatbot service failure and robot linguistic expression type, with patient trust as the dependent variable, by performing a 2 (healthcare chatbot service failure: process failure vs. outcome failure) X 2 (linguistic expression type: need vs. want) ANOVA. The results showed that healthcare chatbot service failure had a significant effect on patient trust (F (1, 518)\u0026thinsp;=\u0026thinsp;19.95, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and the linguistic expression type of the healthcare chatbot had a significant effect on patient trust (F (1, 518)\u0026thinsp;=\u0026thinsp;355.31, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and the interaction between the linguistic expression type of the healthcare chatbot and the healthcare chatbot service failure showed a significant effect on patient trust (F (1, 518)\u0026thinsp;=\u0026thinsp;12.535, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). It can be seen that the interaction effect between linguistic expression type and healthcare chatbot service failure was significant on patient trust, validating H3.(shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e[Insert Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e about here]\u003c/p\u003e\u003cp\u003eThere was a control variable analysis. Based on the study by Kamide et al. (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), we found that patient acceptance to robot may have an impact on patient trust. Therefore, we used patient acceptance as the independent variable and patient trust as the dependent variable, and the results of the experiment showed that patient acceptance did not have a significant effect on patient trust (F (12, 507)\u0026thinsp;=\u0026thinsp;0.745, p\u0026thinsp;=\u0026thinsp;0.707).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec25\" class=\"Section2\"\u003e\u003ch2\u003e6.4 DISCUSSION\u003c/h2\u003e\u003cp\u003eExperiment 3 demonstrated the interaction effect between linguistic expression type and healthcare chatbot service failure on patient trust. The main effect of healthcare chatbot service failure on patient trust was again validated. The results of the experiment pointed out that faced the robot service failure, patient trust was significantly higher in the \u0026ldquo;need\u0026rdquo; group than in the \u0026ldquo;want\u0026rdquo; group. Specifically, with the healthcare chatbot service process failure, the difference between patient trust in the \u0026ldquo;need\u0026rdquo; group and the \u0026ldquo;want\u0026rdquo; group is smaller, showing a tendency to intersect; with the healthcare chatbot service outcome failure, the difference between the two is larger, showing a tendency to diverge. At the same time, we exclude the impact of the patient acceptance toward the robot on the experimental results, which enhance its robustness.\u003c/p\u003e\u003c/div\u003e"},{"header":"7 GENERAL DISCUSSION","content":"\u003cp\u003eWhat kind of message should be sent to conveys important information in a healthcare chatbot service failure scenario. Research has generally found that healthcare chatbot service failure significantly reduces patient trust (Park, \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). However, this study showed that healthcare chatbot attempts to redeem patient trust with messages that help patients re-establish their trust in the robot. While sending out messages is indeed an integral part of successfully building human-robot trust, academics have not yet considered why messages sent by a healthcare chatbot can reshape trust between humans and robots. This study addressed a question that has not been adequately investigated: how robots affect patient trust with service failure.\u003c/p\u003e\u003cp\u003eWe proposed and tested an idea that different semantic frames (need vs. want) in sentences sent by the robot in a healthcare chatbot service failure (process failure vs. outcome failure) scenario affects patients' evaluations for the value of the healthcare chatbot and their subsequent willingness to trust. More specifically, we focused on the significance of patient mindset (growth vs. fixation) and semantic frames (need vs. want) in the healthcare chatbot service failure scenario to enhance patient trust (Su et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The experiment demonstrated that patients do associate healthcare chatbot service failure scenarios with their own mindset, especially when confronted with a robot service process failure. We further showed why the healthcare chatbot semantic framework (need vs. want) can be effective in restoring patients and ultimately enhancing patient trust. This study also indicated the meaningful impact of patient mindset (growth vs. fixed), and robot semantic frames (need vs. want) on regaining patient trust in the face of healthcare chatbot service failure.\u003c/p\u003e\u003cp\u003eWe suggested that in the case of healthcare chatbot service failure, sending more sentences with \"want\" would increase patient trust; conversely, sending sentences with \"need\" would decrease patient trust. However, there was a significant interaction between healthcare chatbot service failure (process failure vs. outcome failure) and patient mindset (growth vs. fixed). Th pattern of this effect adds complexity to the theory of this study. When a healthcare chatbot interacts with patients, they are more likely to be influenced by their own attitudes toward the robot, which ultimately affects patient trust in a healthcare chatbot service failure scenario.\u003c/p\u003e\u003cdiv id=\"Sec27\" class=\"Section2\"\u003e\u003ch2\u003e7.1 THEORETICAL IMPLICATIONS\u003c/h2\u003e\u003cp\u003eAn important contribution that this study has made is to the literature on robot services and patient psychological behaviors, by highlighting that the sent sentences in healthcare chatbot service failure scenarios with different semantic frames, can affect patients' evaluations of the healthcare chatbot and generate different patient trusts. Despite the fact that service robots are increasingly used to improve individual trust in it through verbal interactions (Tussyadiah et al., \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). A large number of studies suggested that the scenarios of robot service failure can salvage individual satisfaction in multiple ways (Belanche et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Liao et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). However, academics have rarely considered the important mechanism of robots influencing patient trust in healthcare chatbot service failure scenarios. As a notable exception, Choi et al. (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) examined efforts made after a healthcare chatbot service failure, but they did not consider the semantic frames in the sent sentences (need vs. want) as well as the patient mindset (growth vs. fixed). Our goal is to bridge the gap between practice and academic research by introducing robust theoretical and experimental approaches to help understand whether patient mindset and semantic frames matter for healthcare chatbot service failure. Our work showed that the question of why and when a healthcare chatbot is able to regain patient trust in the service failure, contributing to the advancement of knowledge on healthcare chatbot service and patient psychological behaviors. Thus, our study unlocked many pathways for academics to learn more about how patient trust can be regained with robot service failure.\u003c/p\u003e\u003cp\u003eThis study also contributed to the development of semantic frames, by integrating semantic frames (need vs. want) and revealing the key mechanisms by which different sent semantic sentences in healthcare chatbot service failure scenarios affect patient trust (deRosset, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). So far, research on semantic frames has only indirectly explored underlying psychological mechanisms (Bless et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e1998\u003c/span\u003e; P. X. Wang et al., \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and has mainly relied on general semantic symbols to perceive scenarios (Beuls et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Therefore, this study has helped to understand the information effect of semantic frame in robot service scenarios through in-depth discussions.\u003c/p\u003e\u003cp\u003eIn addition to extending the scenarios of healthcare chatbot service failure affecting patient trust to everyday services, the theoretical and empirical research also adds to the understanding of patient mindset. As other literature has shown, people's mindset also determines how they perceive robots (Dang \u0026amp; Liu, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The rigorously controlled real-life experiments in this study are equally important in supporting the impact of patient mindset on patient trust with robot service failure.\u003c/p\u003e\u003cp\u003eBased on previous literature on robot service failure, this study demonstrated multiple ways in which service failure can be altered. The experimental results showed that the mechanisms and principles of both process failure and outcome failure in healthcare chatbot service are related to the personal attributes of the patient. These findings can help us to reposition robot service failure from external and uncontrollable factors. Previous studies have found that the factors affecting individual trust lie more in the control of the individual themselves, which helps to redirect the topic of research on robot service failure away from external and uncontrollable factors towards factors of individuals' own control. Although theorists Leo and Huh (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) have hypothesized that compensatory actions in service robot failure scenarios are effective in driving individuals' willingness to adopt service robots again. However, these ideas have so far remained untested by substantial empirical evidence. This study can also address the question of how to regain individuals' trust in healthcare chatbot with service failure.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec28\" class=\"Section2\"\u003e\u003ch2\u003e7.2 PRACTICAL IMPLICATIONS\u003c/h2\u003e\u003cp\u003eThis study demonstrated that the initiative of service robot in sending seemingly irrelevant messages may be beneficial for regaining patients' trust. It is worth noting that we do not believe that the specific discourse of service robots can be the most important means to regain patient trust. However, a relatively simple and feasible discourse for service robots can help to rebuild patient trust in healthcare chatbot service failure scenarios, which is important for healthcare chatbot in real-life activities. Of course, robot discourse is not an intervention in the traditional sense, but something that every healthcare chatbot maker and user should consider. Therefore, this study indicated that an additional set of \"need\" statements for the healthcare chatbot could better compensate for actual healthcare chatbot service failure scenarios, which could have a significant impact on the performance of service robots in the medical institutes.\u003c/p\u003e\u003cp\u003eIn addition, the magnitude of the observed effect sizes can effectively affect the practical implications of the study through its design and variables. They can have a smaller impact on purely scenario-based experimental studies that would normally have a larger significance in practice (Bakan, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1966\u003c/span\u003e[Mohajeri, 2020 #96)]. The patient trust as an explanatory variable implies that patients' daily healthcare activities usually require contact with either a doctor or a healthcare chatbot, indicating the need and urgency for healthcare chatbot service improvement. Therefore, by exploring the types of healthcare chatbot service failure and how the words sent by the healthcare chatbot in these scenarios affect patient trust, we could provide more practical suggestion.\u003c/p\u003e\u003cp\u003eDifferent information can deal with different scenarios, so effectively designing multiple semantic frames in a single human-robot interaction is crucial for users of healthcare chatbots. Past research has demonstrated that semantic frames are more effective in engaging as well as changing the behavioral intentions of individuals (Gao \u0026amp; Zhang, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and semantic frames for healthcare chatbots provide great service value to medical institutes (Taylor, \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). Finding a way to combine the semantic frames and service failure scenarios in the design of a healthcare chatbot is clearly beneficial to both the manufacturer and the user of the healthcare chatbot (medical institutes and patients). The findings suggested that healthcare chatbot manufacturers should adopt different strategies for dealing with healthcare chatbot service failures. For example, when patients facing the process failures, it is advisable to encourage the patients to re-ask the question in a direct tone. When patients facing the outcome failures, it is advisable to acknowledge the occurred mistake and issue a new statement to help the patient to re-ask their questions.\u003c/p\u003e\u003cp\u003eSimilarly, in healthcare chatbot service failure scenarios, considering the patient mindset and regaining the patient trust can be an effective service tool. However, if healthcare chatbot makers want to regain patient trust by encouraging patient tolerance towards the healthcare chatbot, they should consider guiding patients in advance. For example, informing patients that healthcare chatbots are still in a developmental stage and that they need patients' tolerance and patience. Additionally, creating a growth mindset for patients towards healthcare chatbots is also crucial if medical institutes want to improve the scenarios of healthcare chatbot service failure.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec29\" class=\"Section2\"\u003e\u003ch2\u003e7.3 LIMITATIONS AND FUTURE RESEARCH\u003c/h2\u003e\u003cp\u003eAlthough this study provided new insights into patient psychological behaviors and robot service failure, as well as how they can be used to regain patient trust, there are some limitations that need to be addressed in future research. First, we manipulated the patient mindset, but an inclusive growth mindset towards novelty that is widespread in the human world due to the specificity of individual patients. Consequently, the growth mindset may also lead patients to believe that healthcare chatbot service failures are inevitable, as patients with this mindset may believe that all healthcare chatbots need to be tolerated. Future research could consider a more appropriate experimental design to avoid the confounding impact of the growth mindset on individual patients. Second, the manipulations of healthcare chatbot service failures were mainly classified it into process failure and outcome failure. However, in real-life scenario, healthcare chatbot service failures may include more contingencies, such as service lag or delay. Future research could perhaps manipulate service failures to explore the diverse antecedents of robot service failure to restore individual trust. Last, we argued previously that the language sent by the healthcare chatbot in service failure would be directly effective in affecting patients' subsequent behavior. However, the role of an alternative explanation - patients' skepticism towards the robot's language - has not been explored in this study and could be further investigated in future research.\u003c/p\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eCompliance with Ethical Standards\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e\u0026nbsp; The authors declare that they have no conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study received no funding from any organization or individual.\u003c/p\u003e\n\u003cp\u003eMethod\u003c/p\u003e\n\u003cp\u003eAll methods were carried out in accordance with relevant guidelines and regulations.\u003c/p\u003e\n\u003cp\u003eAll experimental protocols were approved by a named institutional and/or licensing committee.\u003c/p\u003e\n\u003cp\u003eInformed consent was obtained from all subjects and/or their legal guardian(s)\u003c/p\u003e\n\u003cp\u003eThis study was reviewed and approved by the Ethics Committee of Huainan Normal University.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAlhaffar, M. B. A., \u0026amp; Janos, S. (2021). Public health consequences after ten years of the Syrian crisis: a literature review. \u003cem\u003eGlobalization and health\u003c/em\u003e,\u003cem\u003e\u0026nbsp;17\u003c/em\u003e(1), 111. https://doi.org/https://doi.org/10.1186/s12992-021-00762-9\u003c/li\u003e\n \u003cli\u003eArikan, E., Altinigne, N., Kuzgun, E., \u0026amp; Okan, M. (2023). May robots be held responsible for service failure and recovery? The role of robot service provider agents\u0026rsquo; human-likeness. \u003cem\u003eJournal of Retailing and Consumer Services\u003c/em\u003e,\u003cem\u003e\u0026nbsp;70\u003c/em\u003e, 103175. https://doi.org/https://doi.org/10.1016/j.jretconser.2022.103175\u003c/li\u003e\n \u003cli\u003eAsan, O., Bayrak, A. E., \u0026amp; Choudhury, A. (2020). Artificial intelligence and human trust in healthcare: focus on clinicians. \u003cem\u003eJournal of Medical Internet Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;22\u003c/em\u003e(6), e15154. https://doi.org/https://doi.org/10.2196/15154\u003c/li\u003e\n \u003cli\u003eBakan, D. (1966). The test of significance in psychological research. \u003cem\u003ePsychological bulletin\u003c/em\u003e,\u003cem\u003e\u0026nbsp;66\u003c/em\u003e(6), 423. https://doi.org/https://doi.org/10.1037/h0020412\u003c/li\u003e\n \u003cli\u003eBaker, D. E. (2018). \u003cem\u003eHow Want vs Need Self-Talk Facilitates Goal-Directed Behavior\u003c/em\u003e. University of Arkansas.\u003c/li\u003e\n \u003cli\u003eBarone, A. M., Stagno, E., \u0026amp; Donato, C. (2024). Call it robot: anthropomorphic framing and failure of self-service technologies. \u003cem\u003eJournal of Services Marketing\u003c/em\u003e,\u003cem\u003e\u0026nbsp;38\u003c/em\u003e(3), 272-287. https://doi.org/https://doi.org/10.1108/JSM-05-2023-0169\u003c/li\u003e\n \u003cli\u003eBecker, M., Mahr, D., \u0026amp; Odekerken-Schr\u0026ouml;der, G. (2023). Customer comfort during service robot interactions. \u003cem\u003eService Business\u003c/em\u003e,\u003cem\u003e\u0026nbsp;17\u003c/em\u003e(1), 137-165. https://doi.org/https://doi.org/10.1007/s11628-022-00499-4\u003c/li\u003e\n \u003cli\u003eBelanche, D., Casal\u0026oacute;, L. V., Flavi\u0026aacute;n, C., \u0026amp; Schepers, J. (2020). Service robot implementation: a theoretical framework and research agenda. \u003cem\u003eThe Service Industries Journal\u003c/em\u003e,\u003cem\u003e\u0026nbsp;40\u003c/em\u003e(3-4), 203-225. https://doi.org/https://doi.org/10.1080/02642069.2019.1672666\u003c/li\u003e\n \u003cli\u003eBeuls, K., Van Eecke, P., \u0026amp; Cangalovic, V. S. (2021). A computational construction grammar approach to semantic frame extraction. \u003cem\u003eLinguistics Vanguard\u003c/em\u003e,\u003cem\u003e\u0026nbsp;7\u003c/em\u003e(1), 20180015. https://doi.org/https://doi.org/10.1515/lingvan-2018-0015\u003c/li\u003e\n \u003cli\u003eBless, H., Betsch, T., \u0026amp; Franzen, A. (1998). Framing the framing effect: The impact of context cues on solutions to the \u0026lsquo;Asian disease\u0026rsquo;problem. \u003cem\u003eEuropean journal of social psychology\u003c/em\u003e,\u003cem\u003e\u0026nbsp;28\u003c/em\u003e(2), 287-291. https://doi.org/https://doi.org/10.1002/(SICI)1099-0992(199803/04)28:2\u0026lt;287::AID-EJSP861\u0026gt;3.0.CO;2-U\u003c/li\u003e\n \u003cli\u003eBock, D. E., Wolter, J. S., \u0026amp; Ferrell, O. (2020). Artificial intelligence: Disrupting what we know about services. \u003cem\u003eJournal of Services Marketing\u003c/em\u003e,\u003cem\u003e\u0026nbsp;34\u003c/em\u003e(3), 317-334. https://doi.org/https://doi.org/10.1108/JSM-01-2019-0047\u003c/li\u003e\n \u003cli\u003eBrooks, R. A. (1991). New approaches to robotics. \u003cem\u003eScience\u003c/em\u003e,\u003cem\u003e\u0026nbsp;253\u003c/em\u003e(5025), 1227-1232. https://doi.org/https://doi.org/10.1126/science.253.5025.1227\u003c/li\u003e\n \u003cli\u003eChao, S., Mu, Q. S., Austin, J., \u0026amp; Mueller, C. M. (2023). Parental Health Mindset Is Related to Preoperative Perceptions of Chest Wall Anomalies. \u003cem\u003eJournal of Surgical Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;284\u003c/em\u003e, 318-321. https://doi.org/https://doi.org/10.1016/j.jss.2022.11.046\u003c/li\u003e\n \u003cli\u003eChi, O. H., Chi, C. G., Gursoy, D., \u0026amp; Nunkoo, R. (2023). Customers\u0026rsquo; acceptance of artificially intelligent service robots: The influence of trust and culture. \u003cem\u003eInternational Journal of Information Management\u003c/em\u003e,\u003cem\u003e\u0026nbsp;70\u003c/em\u003e, 102623. https://doi.org/https://doi.org/10.1016/j.ijinfomgt.2023.102623\u003c/li\u003e\n \u003cli\u003eChing, B. C., Bennett, S. D., Morant, N., Heyman, I., Schleider, J. L., Fifield, K., Allen, S., \u0026amp; Shafran, R. (2023). Growth mindset in young people awaiting treatment in a paediatric mental health service: A mixed methods pilot of a digital single-session intervention. \u003cem\u003eClinical Child Psychology and Psychiatry\u003c/em\u003e,\u003cem\u003e\u0026nbsp;28\u003c/em\u003e(2), 637-653. https://doi.org/https://doi.org/10.1177/13591045221105193\u003c/li\u003e\n \u003cli\u003eChiu, C.-y., Dweck, C. S., Tong, J. Y.-y., \u0026amp; Fu, J. H.-y. (1997). Implicit theories and conceptions of morality. \u003cem\u003eJournal of personality and social psychology\u003c/em\u003e,\u003cem\u003e\u0026nbsp;73\u003c/em\u003e(5), 923. https://doi.org/ https://doi.org/10.1037/0022-3514.73.5.923\u003c/li\u003e\n \u003cli\u003eChoi, S., Mattila, A. S., \u0026amp; Bolton, L. E. (2021). To err is human (-oid): how do consumers react to robot service failure and recovery? \u003cem\u003eJournal of Service Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;24\u003c/em\u003e(3), 354-371. https://doi.org/https://doi.org/10.1177/1094670520978798\u003c/li\u003e\n \u003cli\u003eChuah, S. H.-W., Aw, E. C.-X., \u0026amp; Yee, D. (2021). Unveiling the complexity of consumers\u0026rsquo; intention to use service robots: An fsQCA approach. \u003cem\u003eComputers in Human Behavior\u003c/em\u003e,\u003cem\u003e\u0026nbsp;123\u003c/em\u003e, 106870.\u003c/li\u003e\n \u003cli\u003eCowger, J. A., Grafton, G., \u0026amp; Shah, P. (2020). Avoiding the \u0026ldquo;set it and forget it mentality\u0026rdquo;: a need to regularly reassess left ventricular assist device patients for optimal support. \u003cem\u003eThe Journal of thoracic and cardiovascular surgery\u003c/em\u003e,\u003cem\u003e\u0026nbsp;159\u003c/em\u003e(4), 1322-1325. https://doi.org/https://doi.org/10.1016/j.jtcvs.2019.06.134\u003c/li\u003e\n \u003cli\u003eDang, J., \u0026amp; Liu, L. (2022). A growth mindset about human minds promotes positive responses to intelligent technology. \u003cem\u003eCognition\u003c/em\u003e,\u003cem\u003e\u0026nbsp;220\u003c/em\u003e, 104985. https://doi.org/https://doi.org/10.1016/j.cognition.2021.104985\u003c/li\u003e\n \u003cli\u003edeRosset, L. (2024). A Semantic Framework for the Impure Logic of Ground. \u003cem\u003eJournal of Philosophical Logic\u003c/em\u003e, 1-29. https://doi.org/https://doi.org/10.1007/s10992-023-09738-z\u003c/li\u003e\n \u003cli\u003eDesveaux, L., \u0026amp; Ivers, N. (2024). Practice or perfect? Coaching for a growth mindset to improve the quality of healthcare. \u003cem\u003eBMJ Quality \u0026amp; Safety\u003c/em\u003e. https://doi.org/https://doi.org/10.1136/bmjqs-2023-016456\u003c/li\u003e\n \u003cli\u003eDoupe, P., Faghmous, J., \u0026amp; Basu, S. (2019). Machine learning for health services researchers. \u003cem\u003eValue in Health\u003c/em\u003e,\u003cem\u003e\u0026nbsp;22\u003c/em\u003e(7), 808-815. https://doi.org/https://doi.org/10.1016/j.jval.2019.02.012\u003c/li\u003e\n \u003cli\u003eDupont, P. E., Nelson, B. J., Goldfarb, M., Hannaford, B., Menciassi, A., O\u0026rsquo;Malley, M. K., Simaan, N., Valdastri, P., \u0026amp; Yang, G.-Z. (2021). A decade retrospective of medical robotics research from 2010 to 2020. \u003cem\u003eScience robotics\u003c/em\u003e,\u003cem\u003e\u0026nbsp;6\u003c/em\u003e(60), eabi8017. https://doi.org/https://doi.org/10.1126/scirobotics.abi8017\u003c/li\u003e\n \u003cli\u003eFissell, R. B., Schlundt, D., Cavanaugh, K. L., Mueller, C., Umeukeje, E. M., Nair, D., Wild, M., Chariwala, S., Guide, A., \u0026amp; Stewart, T. (2021). A cross-sectional study of fixed and growth mindset in adult peritoneal dialysis patients. \u003cem\u003eJournal of Patient Experience\u003c/em\u003e,\u003cem\u003e\u0026nbsp;8\u003c/em\u003e, 23743735211055289. https://doi.org/https://doi.org/10.1177/23743735211055289\u003c/li\u003e\n \u003cli\u003eFoster, M. E. (2019). Natural language generation for social robotics: opportunities and challenges. \u003cem\u003ePhilosophical Transactions of the Royal Society B\u003c/em\u003e,\u003cem\u003e\u0026nbsp;374\u003c/em\u003e(1771), 20180027. https://doi.org/https://doi.org/10.1098/rstb.2018.0027\u003c/li\u003e\n \u003cli\u003eGao, Y., \u0026amp; Zhang, Z. (2021). Consensus reaching with non-cooperative behavior management for personalized individual semantics-based social network group decision making. \u003cem\u003eJournal of the Operational Research Society\u003c/em\u003e,\u003cem\u003e\u0026nbsp;73\u003c/em\u003e(11), 2518-2535. https://doi.org/https://doi.org/10.1080/01605682.2021.1997654\u003c/li\u003e\n \u003cli\u003eGibelli, F., Ricci, G., Sirignano, A., Turrina, S., \u0026amp; De Leo, D. (2021). The increasing centrality of robotic technology in the context of nursing care: bioethical implications analyzed through a scoping review approach. \u003cem\u003eJournal of Healthcare Engineering\u003c/em\u003e,\u003cem\u003e\u0026nbsp;2021\u003c/em\u003e. https://doi.org/https://doi.org/10.1155/2021/1478025\u003c/li\u003e\n \u003cli\u003eGrant, A. A. (2023). Regaining the student mindset of patient care. \u003cem\u003eAnnals of Vascular Surgery\u003c/em\u003e. https://doi.org/https://doi.org/10.1016/j.avsg.2023.12.009\u003c/li\u003e\n \u003cli\u003eGuo, C., \u0026amp; Li, H. (2022). Application of 5G network combined with AI robots in personalized nursing in China: A literature review. \u003cem\u003eFrontiers in Public Health\u003c/em\u003e,\u003cem\u003e\u0026nbsp;10\u003c/em\u003e, 948303. https://doi.org/https://doi.org/10.3389/fpubh.2022.948303\u003c/li\u003e\n \u003cli\u003eGuo, Y., Yang, Y., Liu, Y., Li, Q., Cao, F., Feng, M., Wu, H., Li, W., \u0026amp; Kang, Y. (2021). Development status and multilevel classification strategy of medical robots. \u003cem\u003eElectronics\u003c/em\u003e,\u003cem\u003e\u0026nbsp;10\u003c/em\u003e(11), 1278. https://doi.org/https://doi.org/10.3390/electronics10111278\u003c/li\u003e\n \u003cli\u003eHan, J., Davids, J., Ashrafian, H., Darzi, A., Elson, D. S., \u0026amp; Sodergren, M. (2022). A systematic review of robotic surgery: From supervised paradigms to fully autonomous robotic approaches. \u003cem\u003eThe International Journal of Medical Robotics and Computer Assisted Surgery\u003c/em\u003e,\u003cem\u003e\u0026nbsp;18\u003c/em\u003e(2), e2358. https://doi.org/ https://doi.org/10.1002/rcs.2358\u003c/li\u003e\n \u003cli\u003eHolland, J., Kingston, L., McCarthy, C., Armstrong, E., O\u0026rsquo;Dwyer, P., Merz, F., \u0026amp; McConnell, M. (2021). Service robots in the healthcare sector. \u003cem\u003eRobotics\u003c/em\u003e,\u003cem\u003e\u0026nbsp;10\u003c/em\u003e(1), 47. https://doi.org/ https://doi.org/10.3390/robotics10010047\u003c/li\u003e\n \u003cli\u003eHolth\u0026ouml;wer, J., \u0026amp; van Doorn, J. (2023). Robots do not judge: service robots can alleviate embarrassment in service encounters. \u003cem\u003eJournal of the Academy of Marketing Science\u003c/em\u003e,\u003cem\u003e\u0026nbsp;51\u003c/em\u003e(4), 767-784.\u003c/li\u003e\n \u003cli\u003eHolzinger, A., Langs, G., Denk, H., Zatloukal, K., \u0026amp; M\u0026uuml;ller, H. (2019). Causability and explainability of artificial intelligence in medicine. \u003cem\u003eWiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery\u003c/em\u003e,\u003cem\u003e\u0026nbsp;9\u003c/em\u003e(4), e1312. https://doi.org/ https://doi.org/10.1002/widm.1312\u003c/li\u003e\n \u003cli\u003eHu, Y., Min, H., \u0026amp; Su, N. (2021). How sincere is an apology? Recovery satisfaction in a robot service failure context. \u003cem\u003eJournal of Hospitality \u0026amp; Tourism Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;45\u003c/em\u003e(6), 1022-1043. https://doi.org/https://doi.org/10.1177/10963480211011533\u003c/li\u003e\n \u003cli\u003eHuang, R., Li, H., Suomi, R., Li, C., \u0026amp; Peltoniemi, T. (2023). Intelligent physical robots in health care: systematic literature review. \u003cem\u003eJournal of Medical Internet Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;25\u003c/em\u003e, e39786. https://doi.org/https://doi.org/10.2196/39786\u003c/li\u003e\n \u003cli\u003eHwang, J., Park, S., \u0026amp; Kim, I. (2020). Understanding motivated consumer innovativeness in the context of a robotic restaurant: The moderating role of product knowledge. \u003cem\u003eJournal of Hospitality and Tourism Management\u003c/em\u003e,\u003cem\u003e\u0026nbsp;44\u003c/em\u003e, 272-282.\u003c/li\u003e\n \u003cli\u003eJia, Y., Wyer Jr, R. S., \u0026amp; Shen, H. (2021). \u0026ldquo;Will you?\u0026rdquo; versus \u0026ldquo;can you?\u0026rdquo;: Verbal framing moderates the effect of feelings of power on consumers\u0026rsquo; reactions to waiting. \u003cem\u003eJournal of Experimental Psychology: Applied\u003c/em\u003e,\u003cem\u003e\u0026nbsp;27\u003c/em\u003e(2), 213. https://doi.org/https://doi.org/10.1037/xap0000336\u003c/li\u003e\n \u003cli\u003eJin, E., \u0026amp; Eastin, M. S. (2022). When a chatbot smiles at you: The psychological mechanism underlying the effects of friendly language use by product recommendation chatbots. \u003cem\u003eCyberpsychology, Behavior, and Social Networking\u003c/em\u003e,\u003cem\u003e\u0026nbsp;25\u003c/em\u003e(9), 597-604. https://doi.org/https://doi.org/10.1089/cyber.2021.0318\u003c/li\u003e\n \u003cli\u003eJoseph, G. P., Segovia, N. A., Wright, R. C., Mueller, C., \u0026amp; Tileston, K. R. (2021). Mindset correlates with health-related quality of life assessment in patients with adolescent idiopathic scoliosis. \u003cem\u003eSpine Deformity\u003c/em\u003e,\u003cem\u003e\u0026nbsp;9\u003c/em\u003e, 349-354. https://doi.org/https://doi.org/10.1007/s43390-020-00243-w\u003c/li\u003e\n \u003cli\u003eKakhi, K., Alizadehsani, R., Kabir, H. D., Khosravi, A., Nahavandi, S., \u0026amp; Acharya, U. R. (2022). The internet of medical things and artificial intelligence: trends, challenges, and opportunities. \u003cem\u003eBiocybernetics and Biomedical Engineering\u003c/em\u003e,\u003cem\u003e\u0026nbsp;42\u003c/em\u003e(3), 749-771. https://doi.org/https://doi.org/10.1016/j.bbe.2022.05.008\u003c/li\u003e\n \u003cli\u003eKamide, H., Kawabe, K., Shigemi, S., \u0026amp; Arai, T. (2014). Relationship between familiarity and humanness of robots\u0026ndash;quantification of psychological impressions toward humanoid robots. \u003cem\u003eAdvanced Robotics\u003c/em\u003e,\u003cem\u003e\u0026nbsp;28\u003c/em\u003e(12), 821-832. https://doi.org/https://doi.org/10.1080/01691864.2014.893837\u003c/li\u003e\n \u003cli\u003eKerasidou, A. (2020). Artificial intelligence and the ongoing need for empathy, compassion and trust in healthcare. \u003cem\u003eBulletin of the World Health Organization\u003c/em\u003e,\u003cem\u003e\u0026nbsp;98\u003c/em\u003e(4), 245. https://doi.org/https://doi.org/10.2471/BLT.19.237198\u003c/li\u003e\n \u003cli\u003eKhaksar, S. M. S., Shahmehr, F. S., Miah, S., Daim, T., \u0026amp; Ozdemir, D. (2024). Privacy concerns versus personalisation benefits in social robot acceptance by employees: A paradox theory\u0026mdash;Contingency perspective. \u003cem\u003eTechnological Forecasting and Social Change\u003c/em\u003e,\u003cem\u003e\u0026nbsp;198\u003c/em\u003e, 123034. https://doi.org/https://doi.org/10.1016/j.techfore.2023.123034\u003c/li\u003e\n \u003cli\u003eKoutentakis, D., Pilozzi, A., \u0026amp; Huang, X. (2020). Designing socially assistive robots for Alzheimer\u0026rsquo;s disease and related dementia patients and their caregivers: where we are and where we are headed. Healthcare,\u003c/li\u003e\n \u003cli\u003eKuchenbrandt, D., Eyssel, F., Bobinger, S., \u0026amp; Neufeld, M. (2013). When a robot\u0026rsquo;s group membership matters: Anthropomorphization of robots as a function of social categorization. \u003cem\u003eInternational Journal of Social Robotics\u003c/em\u003e,\u003cem\u003e\u0026nbsp;5\u003c/em\u003e, 409-417. https://doi.org/https://doi.org/10.1007/s12369-013-0197-8\u003c/li\u003e\n \u003cli\u003eLee, H., Piao, M., Lee, J., Byun, A., \u0026amp; Kim, J. (2020). The purpose of bedside robots: exploring the needs of inpatients and healthcare professionals. \u003cem\u003eCIN: Computers, Informatics, Nursing\u003c/em\u003e,\u003cem\u003e\u0026nbsp;38\u003c/em\u003e(1), 8-17. https://doi.org/https://doi.org/10.1097/CIN.0000000000000558\u003c/li\u003e\n \u003cli\u003eLeo, X., \u0026amp; Huh, Y. E. (2020). Who gets the blame for service failures? Attribution of responsibility toward robot versus human service providers and service firms. \u003cem\u003eComputers in Human Behavior\u003c/em\u003e,\u003cem\u003e\u0026nbsp;113\u003c/em\u003e, 106520. https://doi.org/https://doi.org/10.1016/j.chb.2020.106520\u003c/li\u003e\n \u003cli\u003eLiao, G.-Y., Huang, T.-L., Wong, M.-K., Shyu, Y.-I. L., Ho, L.-H., Wang, C., Cheng, T., \u0026amp; Teng, C.-I. (2023). Enhancing Nurse\u0026ndash;Robot Engagement: Two-Wave Survey Study. \u003cem\u003eJournal of Medical Internet Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;25\u003c/em\u003e, e37731. https://doi.org/https://doi.org/10.2196/37731\u003c/li\u003e\n \u003cli\u003eLiu, D., Li, C., Zhang, J., \u0026amp; Huang, W. (2023). Robot service failure and recovery: Literature review and future directions. \u003cem\u003eInternational Journal of Advanced Robotic Systems\u003c/em\u003e,\u003cem\u003e\u0026nbsp;20\u003c/em\u003e(4), 17298806231191606. https://doi.org/https://doi.org/10.1177/17298806231191606\u003c/li\u003e\n \u003cli\u003eLiu, J., Kong, X., Xia, F., Bai, X., Wang, L., Qing, Q., \u0026amp; Lee, I. (2018). Artificial intelligence in the 21st century. \u003cem\u003eIeee Access\u003c/em\u003e,\u003cem\u003e\u0026nbsp;6\u003c/em\u003e, 34403-34421. https://doi.org/https://doi.org/10.1109/ACCESS.2018.2819688\u003c/li\u003e\n \u003cli\u003eLiu, X., He, X., Wang, M., \u0026amp; Shen, H. (2022). What influences patients\u0026apos; continuance intention to use AI-powered service robots at hospitals? The role of individual characteristics. \u003cem\u003eTechnology in Society\u003c/em\u003e,\u003cem\u003e\u0026nbsp;70\u003c/em\u003e, 101996. https://doi.org/https://doi.org/10.1016/j.techsoc.2022.101996\u003c/li\u003e\n \u003cli\u003eLuo, Z., Cheng, W., Zhao, T., \u0026amp; Xiang, N. (2024). Intelligent sensory systems toward soft robotics. \u003cem\u003eApplied Materials Today\u003c/em\u003e,\u003cem\u003e\u0026nbsp;37\u003c/em\u003e, 102122. https://doi.org/https://doi.org/10.1016/j.apmt.2024.102122\u003c/li\u003e\n \u003cli\u003eMende, M., Scott, M. L., Van Doorn, J., Grewal, D., \u0026amp; Shanks, I. (2019). Service robots rising: How humanoid robots influence service experiences and elicit compensatory consumer responses. \u003cem\u003eJournal of Marketing Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;56\u003c/em\u003e(4), 535-556. https://doi.org/https://doi.org/10.1177/0022243718822827\u003c/li\u003e\n \u003cli\u003eMeng, Y.-J., Li, N.-X., Liu, C.-J., Chen, J.-H., Song, Y.-C., \u0026amp; Qian, Z.-S. (2008). Quality of life and hostile mentality trend of patients with HIV/AIDS in China. \u003cem\u003ePublic Health\u003c/em\u003e,\u003cem\u003e\u0026nbsp;122\u003c/em\u003e(4), 404-411. https://doi.org/https://doi.org/10.1016/j.puhe.2007.06.007\u003c/li\u003e\n \u003cli\u003eMoriuchi, E., \u0026amp; Murdy, S. (2024). The role of robots in the service industry: Factors affecting human-robot interactions. \u003cem\u003eInternational Journal of Hospitality Management\u003c/em\u003e,\u003cem\u003e\u0026nbsp;118\u003c/em\u003e, 103682. https://doi.org/https://doi.org/10.1016/j.ijhm.2023.103682\u003c/li\u003e\n \u003cli\u003ePackard, G., \u0026amp; Berger, J. (2017). How language shapes word of mouth\u0026apos;s impact. \u003cem\u003eJournal of Marketing Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;54\u003c/em\u003e(4), 572-588. https://doi.org/https://doi.org/10.1509/jmr.15.0248\u003c/li\u003e\n \u003cli\u003ePark, S. (2020). Multifaceted trust in tourism service robots. \u003cem\u003eAnnals of Tourism Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;81\u003c/em\u003e, 102888. https://doi.org/https://doi.org/10.1016/j.annals.2020.102888\u003c/li\u003e\n \u003cli\u003ePennebaker, J. W., Mehl, M. R., \u0026amp; Niederhoffer, K. G. (2003). Psychological aspects of natural language use: Our words, our selves. \u003cem\u003eAnnual review of psychology\u003c/em\u003e,\u003cem\u003e\u0026nbsp;54\u003c/em\u003e(1), 547-577. https://doi.org/https://doi.org/10.1146/annurev.psych.54.101601.145041\u003c/li\u003e\n \u003cli\u003ePollitt, J. (1982). Moodiness: a heavenly problem? Observations on the pathology of moodiness. \u003cem\u003eJournal of the Royal Society of Medicine\u003c/em\u003e,\u003cem\u003e\u0026nbsp;75\u003c/em\u003e(1), 7-16. https://doi.org/https://doi.org/10.1177/014107688207500104\u003c/li\u003e\n \u003cli\u003eQian, K., \u0026amp; Yamada, Y. (2020). Exploring the role of the behavioral immune system in acceptability of entomophagy using semantic associations and food-related attitudes. \u003cem\u003eFrontiers in Nutrition\u003c/em\u003e,\u003cem\u003e\u0026nbsp;7\u003c/em\u003e, 66. https://doi.org/https://doi.org/10.3389/fnut.2020.00066\u003c/li\u003e\n \u003cli\u003eReich, T., Kaju, A., \u0026amp; Maglio, S. J. (2023). How to overcome algorithm aversion: Learning from mistakes. \u003cem\u003eJournal of Consumer Psychology\u003c/em\u003e,\u003cem\u003e\u0026nbsp;33\u003c/em\u003e(2), 285-302. https://doi.org/ https://doi.org/10.1002/jcpy.1313\u003c/li\u003e\n \u003cli\u003eRichardson, D., Kinnear, B., Hauer, K. E., Turner, T. L., Warm, E. J., Hall, A. K., Ross, S., Thoma, B., Van Melle, E., \u0026amp; Collaborators, I. (2021). Growth mindset in competency-based medical education. \u003cem\u003eMedical Teacher\u003c/em\u003e,\u003cem\u003e\u0026nbsp;43\u003c/em\u003e(7), 751-757. https://doi.org/https://doi.org/10.1080/0142159X.2021.1928036\u003c/li\u003e\n \u003cli\u003eRoesler, E., Heuring, M., \u0026amp; Onnasch, L. (2023). (hu) man-like robots: The impact of anthropomorphism and language on perceived robot gender. \u003cem\u003eInternational Journal of Social Robotics\u003c/em\u003e,\u003cem\u003e\u0026nbsp;15\u003c/em\u003e(11), 1829-1840. https://doi.org/https://doi.org/10.1007/s12369-023-00975-5\u003c/li\u003e\n \u003cli\u003eRyoo, Y., Jeon, Y. A., \u0026amp; Kim, W. (2024). The blame shift: Robot service failures hold service firms more accountable. \u003cem\u003eJournal of Business Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;171\u003c/em\u003e, 114360. https://doi.org/https://doi.org/10.1016/j.jbusres.2023.114360\u003c/li\u003e\n \u003cli\u003eSchraft, R., Degenhart, E., Hagele, M., \u0026amp; Kahmeyer, M. (1993). New robot applications in production and service. Proceedings of 1993 IEEE/Tsukuba International Workshop on Advanced Robotics,\u003c/li\u003e\n \u003cli\u003eSeo, H., Park, J., Bennis, M., \u0026amp; Debbah, M. (2023). Semantics-native communication via contextual reasoning. \u003cem\u003eIEEE Transactions on Cognitive Communications and Networking\u003c/em\u003e. https://doi.org/10.1109/TCCN.2023.3250206\u003c/li\u003e\n \u003cli\u003eShanks, I., Scott, M. L., Mende, M., van Doorn, J., \u0026amp; Grewal, D. (2024). Cobotic service teams and power dynamics: Understanding and mitigating unintended consequences of human-robot collaboration in healthcare services. \u003cem\u003eJournal of the Academy of Marketing Science\u003c/em\u003e, 1-27. https://doi.org/https://doi.org/10.1007/s11747-024-01004-1\u003c/li\u003e\n \u003cli\u003eSu, L., Sengupta, J., Li, Y., \u0026amp; Chen, F. (2024). \u0026ldquo;Want\u0026rdquo; versus \u0026ldquo;Need\u0026rdquo;: How Linguistic Framing Influences Responses to Crowdfunding Appeals. \u003cem\u003eJournal of Consumer Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;50\u003c/em\u003e(5), 923-944. https://doi.org/https://doi.org/10.1093/jcr/ucad033\u003c/li\u003e\n \u003cli\u003eTaylor, R. H. (2006). A perspective on medical robotics. \u003cem\u003eProceedings of the IEEE\u003c/em\u003e,\u003cem\u003e\u0026nbsp;94\u003c/em\u003e(9), 1652-1664. https://doi.org/https://doi.org/10.1109/JPROC.2006.880669\u003c/li\u003e\n \u003cli\u003eTorrent-Sellens, J., Jim\u0026eacute;nez-Zarco, A. I., \u0026amp; Saig\u0026iacute;-Rubi\u0026oacute;, F. (2021). Do people trust in robot-assisted surgery? Evidence from Europe. \u003cem\u003eInternational Journal of Environmental Research and Public Health\u003c/em\u003e,\u003cem\u003e\u0026nbsp;18\u003c/em\u003e(23), 12519. https://doi.org/https://doi.org/10.3390/ijerph182312519\u003c/li\u003e\n \u003cli\u003eTussyadiah, I. P., Zach, F. J., \u0026amp; Wang, J. (2020). Do travelers trust intelligent service robots? \u003cem\u003eAnnals of Tourism Research\u003c/em\u003e,\u003cem\u003e\u0026nbsp;81\u003c/em\u003e, 102886. https://doi.org/https://doi.org/10.1016/j.annals.2020.102886\u003c/li\u003e\n \u003cli\u003eWang, P. X., Wang, Y., \u0026amp; Jiang, Y. (2023). Gift or donation? Increase the effectiveness of charitable solicitation through framing charitable giving as a gift. \u003cem\u003eJournal of Marketing\u003c/em\u003e,\u003cem\u003e\u0026nbsp;87\u003c/em\u003e(1), 133-147. https://doi.org/https://doi.org/10.1177/00222429221081506\u003c/li\u003e\n \u003cli\u003eWang, X., Hwang, Y., \u0026amp; Guchait, P. (2023). When robot (vs. human) employees say \u0026ldquo;sorry\u0026rdquo; following service failure. \u003cem\u003eInternational Journal of Hospitality \u0026amp; Tourism Administration\u003c/em\u003e,\u003cem\u003e\u0026nbsp;24\u003c/em\u003e(4), 540-562. https://doi.org/https://doi.org/10.1080/15256480.2021.2017812\u003c/li\u003e\n \u003cli\u003eWatson, D., Clark, L. A., \u0026amp; Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. \u003cem\u003eJournal of personality and social psychology\u003c/em\u003e,\u003cem\u003e\u0026nbsp;54\u003c/em\u003e(6), 1063. https://doi.org/https://doi.org/10.1037/0022-3514.54.6.1063\u003c/li\u003e\n \u003cli\u003eZebrowski, R. L. (2020). Fear of a bot planet: Anthropomorphism, humanoid embodiment, and machine consciousness. \u003cem\u003eJournal of Artificial Intelligence and Consciousness\u003c/em\u003e,\u003cem\u003e\u0026nbsp;7\u003c/em\u003e(01), 119-132. https://doi.org/https://doi.org/10.1142/S2705078520500071\u003c/li\u003e\n \u003cli\u003eZhang, M., Cui, J., \u0026amp; Zhong, J. (2023). How consumers react differently toward humanoid vs. nonhumanoid robots after service failures: a moderated chain mediation model. \u003cem\u003eInternational Journal of Emerging Markets\u003c/em\u003e. https://doi.org/https://doi.org/10.1108/IJOEM-06-2022-1023\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"healthcare chatbot, patient trust, service failure, linguistic expression type","lastPublishedDoi":"10.21203/rs.3.rs-6930952/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6930952/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study aimed to investigate the impact of healthcare chatbot service failure on patient trust and examined the role of two factors, patient mindset and the linguistic expression type of the healthcare chatbot service. By developing a theoretical framework and conducting the evidence-based experiment, this study argued that healthcare chatbot service failure reduces patient trust in the robot, by affecting patient mindset, as well as changing patients' views and perceptions of the healthcare chatbot service. This study first presented the impact of healthcare chatbot service failure on patients' trust shaping, and explored the interaction effect on patients' trust, from the perspectives of patients' mindset and the linguistic expression type of the healthcare chatbot service. The contribution of this study is to fill the research gap on the impact of the type of healthcare chatbot service failure (process failure vs. outcome failure) on patient trust, and to explore the interaction effect between patient mindset and the linguistic expression type of the healthcare chatbot service and its operational mechanisms. These findings are important for the understanding and application of healthcare chatbot service and patient trust, providing valuable guidance for the development of healthcare chatbot service models, the maintenance of doctor-patient relationships, and service capabilities of hospital.\u003c/p\u003e","manuscriptTitle":"How Healthcare Chatbot Service Failures Affect Patient Trust?","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-16 12:39:02","doi":"10.21203/rs.3.rs-6930952/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4167492a-3889-47ee-866a-a4616a2fbe3a","owner":[],"postedDate":"September 16th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":54530180,"name":"Biological sciences/Psychology"},{"id":54530181,"name":"Biological sciences/Psychology/Human behaviour"}],"tags":[],"updatedAt":"2026-01-05T06:10:09+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-16 12:39:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6930952","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6930952","identity":"rs-6930952","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}