Inequity Aversion Toward AI Counterparts

Inequity Aversion Toward AI Counterparts | 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Article Inequity Aversion Toward AI Counterparts Debanjan Borthakur, Peter Diep, Jason Plaks This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6051145/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 Human moral interactions often assume that resources should be allocated equitably, i.e., one should not take more than one’s fair share. To what extent do people apply this assumption to social AI entities? Using a 21-round Ultimatum Game, we investigated participants’ behavioral, physiological, and affective responses to fair, disadvantageous, and advantageous offers from an AI (vs. human) counterpart. We report three principal findings: (a) Participants were more likely to reject disadvantageous offers from an AI counterpart than from a human counterpart, but were more likely to reject advantageous offers from a human counterpart than from an AI counterpart; (b) Participants reported more negative affect following disadvantageous offers from an AI counterpart than from a human counterpart; (c) Participants exhibited a stronger association between heart rate variability and rejection rate for disadvantageous offers from an AI counterpart than from a human counterpart. Based on these findings, we propose a model emphasizing an important, previously under-examined role of self-regulatory processes in humans’ responses toward AI moral behavior. Social science/Psychology Biological sciences/Psychology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction The implicit or explicit assumption of fair treatment is a cornerstone of human moral behavior (Pettit, 2018 ; Sloane, Baillargeon, & Premack, 2012 ). Numerous studies have provided evidence that individuals are more likely to behave prosocially to the extent that they trust that others will allocate resources equitably and will not systematically abuse such trust (e.g., Nowak & Sigmund, 2005 ; Sheldon et al., 2018 ). The research literatures on resource allocation and on interpersonal trust have, understandably, focused almost exclusively on human-to-human interactions (e.g., Axelrod, 2006 ; Evans & Revelle, 2008 ; Kosfeld et al., 2005 ; Rotter, 1967 ). However, with the increasing capabilities of LLM-based social agents to simulate human behavior, a growing literature has begun to turn these questions toward humans’ moral expectations of AI agents with the aim of better understanding the mechanisms - and limits - of anthropomorphism (e.g., Karpus et al. 2021 ; Nielsen et al., 2022b; Oudah et al., 2024 ; Malle & Ullman, 2021 ; Plaks, Bustos-Rodriguez & Ayad, 2022). Several research teams have approached these questions using classic behavioral economic paradigms 1 . For example, Karpus et al. ( 2021 ) reported that in a one-shot Prisoner’s Dilemma game, participants expected the same degree of cooperative behavior from an AI counterpart as from a human, although they themselves responded significantly more cooperatively toward the human (49%) than toward the AI counterpart (36%). (See also Plaks et al., 2022 .) Similarly, in a one-shot Trust Game, participants expected equivalent payouts from a human or an AI counterpart, although they themselves reciprocated prosocially 75% of the time when their counterpart was human, but only 34% when it was an AI agent. Karpus et al. ( 2021 ) reported analogous results with a one-shot Dictator Game and Nielsen et al. (2022b) reported parallel findings with a Public Goods game. In summary, although participants tend to approach an AI counterpart with the assumption that it will treat them as equitably as a human will, they feel less obligated to respond in kind toward the AI counterpart (Bonnefon, Rahwan, & Shariff, 2024). The Ultimatum Game Although the Prisoner’s Dilemma, Trust Game, and Dictator Game all partially concern expectations of fair treatment, they also invoke a range of further concerns, including interpersonal trust and strategic gamesmanship. Arguably a more direct method to assess assumptions of fair treatment per se is the Ultimatum Game (UG) (Güth et al., 1982 ; Harsanyi, 1961 ). In a one-shot UG, the Proposer starts with a sum of money. The Proposer decides how to split the money with the Responder (who is made aware of the total sum of money). The Responder may accept or reject this split. If the Responder accepts it, the money is split accordingly. If the Responder rejects the split, both players receive nothing. Before play begins, both players are informed of all possible outcomes. Thus, if the Responder rejects a disadvantageous split (e.g., 20%-80%) - thereby choosing to receive nothing rather than something - this can be interpreted as a signal of the Responder’s displeasure over the violation of an implied expectation of fair treatment. Studies using the UG have consistently revealed that, indeed, a higher proportion of Responders reject disadvantageous splits (typically starting around the 30%Responder-70%Proposer range) than would be predicted by orthodox, rational choice theories (Nowak et al., 2000 ; Morewedge et al., 2014 ). In this respect, the UG differs from the related Dictator Game (DG); in the DG, the Responder is not afforded the opportunity to reject the offer. To what extent does this pattern extend to human-AI interaction? In one of the first studies to investigate the UG in this context, Sanfey et al. (2003) reported from a small sample (19 participants) that Responders reacted with more emotion (likely anger) in response to a disadvantageous offer from a human than from a computer counterpart. Participants’ anterior insula activation during the rejection of inequitable offers correlated with emotional expressions, suggesting a possible neural substrate for the relevant emotional processes. The authors proposed that participants reacted more negatively toward the human’s disadvantageous offer than the computer’s disadvantageous offer because people grant a higher degree of autonomy (Plaks et al., 2022 ), agency (Oudah et al., 2024 ), intentionality (Ayad & Plaks, 2025), and, in turn, moral responsibility (e.g. Strohminger & Jordan, 2022 ) to humans than to AI agents. (See also van 't Wout et al., 2006 for related findings.) Evolving Expectations of AI Agents In the intervening decades, however, enormous advances in computational power and neural architecture have led to the proliferation of commercially available AI products. This has brought human-AI interaction more squarely into the realm of everyday experience. As such, expectations about AI moral behavior are likely to have evolved. The idea of anthropomorphically blaming, for example, ChatGPT (not the programmers, not the OpenAI corporation) for a perceived transgression may seem less far-fetched to laypeople in 2025 than in the past. For this reason, we hypothesized a reversal of the Sanfey et al. effect in the Ultimatum Game : When provided with a disadvantageous offer, present-day people will respond more negatively toward AI agents than toward humans. We suspected this for four reasons. First, recent studies have demonstrated that people generally hold machines (relative to humans) to - and expect - a higher standard of reliability, a phenomenon dubbed the Perfect Automation Schema (Merritt et al., 2015 ; see also Oudah et al., 2024 ; Shariff et al., 2021 ). Thus, a violation of such expected reliability and fairness represents a greater betrayal (Bohnet & Zeckhauser, 2004 ). Second, many people may hold the assumption that AI agents should serve humans, not undermine them (e.g., de Freitas et al., 2023 ). A disadvantageous offer may violate the AI agent’s presumed subordinate role. Third, interpersonal politeness norms that demand negative emotion suppression (Smith, Young, & Ford, 2023 ) may be reduced or absent when interacting with an AI counterpart (compared to a human counterpart). That is, people may feel freer to express their outrage toward an entity that is assumed to lack the capacity to experience feelings, such as insult or hurt (Oudah et al., 2024 ). Fourth, participants may feel less reluctant to aggress toward, or even ‘punish’, an AI counterpart because they see it as a mechanism that possesses fewer (if any) moral rights (Strohminger & Jordan, 2022 ; Treiman et al., 2023 ). Advantageous Offers Are Also a Violation of Fairness Norms Considerably less work in this area has examined the psychology of receiving an unfairly advantageous offer, or advantageous inequity . Our method permitted us to measure participants’ offer rejection rate, self-reported affect, and heart-rate variability in response to the experience of being over benefitted (e.g., 80%Responder-20%Proposer) by either a human or an AI counterpart. While a strict, rational self-interest approach might suggest that people should enjoy receiving unexpectedly high returns, there are grounds to suspect that, in practice, people will generally find advantageous inequity discomfiting. First, people may experience feelings of guilt upon receiving a payout that is perceived to be larger than deserved. This concept has a long history in psychological research, including classic work in Equity Theory in social relationships (Walster, 1978), as well as relatively recent conceptualizations (Bolton & Ockenfels, 2000 ; Dunning et al., 2014 ). Second, people may doubt their ability to maintain the positive outcomes into the future and may thus experience anxiety over the likely impending decrease (e.g., Plaks & Stecher, 2007). Third, separately from the valence of the payout (i.e., over- vs. under benefit), people may generally experience confusion upon receiving any sort of unexpected result. Finally, the literature on downregulation of positive emotion (e.g., Schall et al., 2016 ; Zou, Plaks, & Peterson, 2019 ) indicates that, upon receiving positive news, people are typically sensitive to the emotional state of others in their environment and calibrate their own emotional expression so as not to deviate too much in a positive direction. Studies on Advantageous Inequity Aversion (AIA) have demonstrated that medial frontal negativity (MFN), an index of neural activity associated with expectancy violation, responds negatively to advantageous inequities (Oliveira et al., 2007 ; Wu and Zhou, 2009 ). This aversion to unfair advantage aligns with societal norms that equate fairness with equal treatment, an observation supported by behavioral responses in the Ultimatum Game, in which advantageous offers are often rejected despite the potential gain (McAuliffe et al., 2014 ). Furthermore, research with children indicates that this aversion is socially conditioned and develops with age. Whereas younger children tend to accept unfair advantages, by around age 8 they begin to reject such benefits, prioritizing fairness over personal gain (Blake & McAuliffe, 2011 ; McAuliffe et al., 2013 ). Finally, there is evidence that people tend to view inequities (whether advantageous or disadvantageous) as violations of moral principles (Shaw & Choshen-Hillel, 2017 ). To what extent might such concerns extend to AI agents? Given that most people assume that AI entities are (a) largely incapable of experiencing emotions (Oudah et al., 2024 ) and (b) less socially embedded than humans (Ayad & Plaks, 2025), these interpersonal emotional concerns should be largely moot when one is over benefitted by an AI agent. Thus, we hypothesized that participants would display lower advantageous inequity aversion (expressed by lower rejection rates, lower emotional disturbance, and higher heart-rate variability) when over benefitted by an AI counterpart than when over benefitted by a human counterpart. Our design involved a 21-round game, rather than one shot. This allowed us to aggregate across participants’ responses to multiple fair or unfair offers for arguably a more robust measure than can be provided by a single, one-shot game. In addition, the extended play period permitted us to measure physiological indices of heart rate variability. We suspected that such indices, which we describe next, would provide additional insight into the processes that contribute to accept/reject decisions. Heart Rate Variability A healthy heart does not beat like a regular clock. Instead, its oscillations are complex and non-linear. Heart rate variability (HRV) refers to the fluctuation in the time interval between adjacent heartbeats (McCraty et al., 2015). It indexes neurocardiac function and is generated by heart-brain interactions and dynamic, non-linear autonomic nervous system (ANS) processes (Shaffer et al., 2017). The beat-to-beat fluctuations of a healthy heart are best described by mathematical chaos (Goldberger, 1991 ). Such nonlinear variability is thought to provide the flexibility to rapidly cope with an uncertain and changing environment (Beckers et al., 2006 ). An optimal level of HRV is associated with health, self-regulatory capacity, and adaptability or resilience (McCraty et al., 2015). HRV (measured via time, frequency, and non-linear metrics) has been identified as a physiological marker of stress-induced elevation in the sympathetic nervous system (SNS) and reduction in the parasympathetic nervous system (PNS) (Kim et al., 2018 ). Time-domain indices of HRV quantify the amount of variability in measurements of the inter-beat interval (IBI). Frequency-domain measurements estimate the distribution of absolute or relative power into frequency bands. Non-linear measurements allow researchers to quantify the unpredictability of a time series (Stein & Reddy, 2005 ). Several studies have focused on how individual differences in heart rate variability are associated with decision-making in the UG. For example, Sütterlin et al. ( 2011a ) reported that participants with higher resting HRV, which indicates greater parasympathetic nervous system activity and inhibitory control, were more likely to reject unfair offers in the UG. Additionally, performance on a separate motor response inhibition task (Stop Signal Task, SST) also predicted rejection rates in the UG. Combining HRV and SST measures explained a significant portion of the variance in rejection rates, suggesting that self-regulatory capacity plays a crucial role in overcoming economic self-interest and promoting fairness-related behavior in the UG. In a related study, Dulleck et al. ( 2011 ) reported that lower HRV (indicative of higher self-regulatory effort and/or stress) correlated with both rejecting and making unfair offers in the UG. The primary time-domain measure used to estimate vagally mediated changes reflected in HRV is the root mean square of successive differences between normal heartbeats (RMSSD) (Shaffer et al., 2017). Osumi and Ohira ( 2009 ) found that heart rate deceleration was more pronounced when participants received offers that they subsequently rejected, suggesting that the perception of disadvantage triggers physiological responses that contribute to the decision to reject. Another research team reported that criminal judges were not only slower in their responses but also rejected a greater proportion of unfair offers, with their decision-making and rejection rates significantly correlated with higher HRV scores (Santamaría-García, H. et al., 2021). These data suggest that HRV reflects self-regulatory effort in moral decision-making contexts, including the UG context. Taken together, these findings have introduced a new perspective to UG behavior by highlighting inhibitory control processes that may contribute to the decision to reject an offer. Materials and Methods Participants One hundred and fifty participants initially enrolled in the study. However, due to participant nonresponse or system malfunctions, the final sample size for analysis was 142 participants (Mean age = 19.22, SD = 2.68) 2 , with 63 randomly assigned to the human counterpart condition and 79 to the AI counterpart condition. The sample size was determined using a priori power analysis, informed by effect sizes and sample sizes reported in comparable prior studies (e.g., Sütterlin et al., 2011; Santamaría-García et al., 2021). The study was approved by the Research Ethics Board of the University of Toronto. All participants provided informed consent before the experimental procedures. Socio-demographic data (age, race/ethnicity, socioeconomic status) were collected via end questionnaires. Exclusion criteria included cardiovascular disease, arrhythmias, serious medical conditions, and the use of heart rate variability-affecting medications (e.g., beta-blockers, calcium channel blockers). Participants also reported their frequency of alcohol intake, tobacco intake, caffeine intake, minimum fifteen-minute exercise intervals, and minimum fifteen-minute meditation intervals on a weekly basis, due to their impact on HRV. Participants were required to abstain from smoking and caffeine on the day of the experiment and to avoid strenuous exercise and heavy meals prior to testing. The data supporting the findings of this study are available upon request from the lead author. Procedure After providing informed consent, proposers and responders briefly introduced themselves. In the AI condition, participants were greeted by "SAM," embodied as a small, disk-shaped audio speaker. In reality, a research assistant fulfilled this role remotely, via the ‘Wizard of Oz’ method; Unbeknownst to participants, SAM’s utterances were pre-recorded using Amazon Polly’s text-to-voice function and the sound files were played through the speaker by an assistant in an adjoining room. Thus, we wish to note that although we use the term ‘AI’ to refer to SAM, it was not, in fact, an artificially intelligent agent. However, post-study queries revealed that all participants reported that they believed SAM to be an AI agent. SAM greeted the participants with, “Hi, my name is SAM, it’s nice to meet you today, how have you been doing”? SAM was always presented as the proposer. In the Human condition, the proposer - a research assistant posing as another participant - made a similar verbal greeting. Next, participants completed a series of questionnaires, most of which test hypotheses that are beyond the scope of the present article and will thus not be discussed further and can be found in the supplementary material 3 . Then each participant was taken to a separate room and connected to the physiological recording system. Throughout the session, continuous heart rate monitoring was conducted using PPG sensors. (See details below.) After tests of the PPG system, participants turned their attention to the computer terminal to play a modified version of the Ultimatum Game (UG). To briefly recap the UG procedure, on each round, the proposer offers a split of points ranging from 0 to 100 (e.g., 70/30, such that the proposer keeps 70 and 30 goes to the responder). If the responder (the participant) agrees to the division, both parties receive the proposed amounts. However, a rejection by the responder results in neither player earning any points. All participants were assigned the role of the responder after completing prerequisite comprehension questions about the game rules. Research assistants, posing as fellow participants, posed as proposers in the human counterpart condition. The sequence of offers was quasi-randomized to ensure that each participant received an equal distribution of offers—7 fair, 7 disadvantageous, and 7 advantageous. On each round, the proposer’s offer was presented to participants on the screen. Participants’ task was to select ‘Accept’ or ‘Reject’. The Ultimatum Game was conducted using the oTree platform (Chen, Schonger, & Wickens, 2016 ). Following each of the 21 rounds, participants completed a brief, self-reported affect questionnaire - modeled on Osumi and Ohira ( 2009 ). In keeping with the exact language used by Osumi and Ohira, participants rated their feelings of anger , aversion , reassurance , and pleasure . In addition, we added a new exploratory item ( disgust ) in light of the extensive literature on the link between disgust and moral condemnation (e.g., Inbar, Pizarro, & Bloom 2009 ; Robinson, Xu, & Plaks, 2019 ). Heart Rate Variability Measurement To measure heart rate variability, we used photoplethysmogram (PPG), which detects cardiovascular pulse waves via a light source and detector. This technique captures variations in blood volume or flow, with the intensity of backscattered light corresponding to changes in blood volume (Evans and Geddes 1988 ; Alnaeb et al. 2007 ). PPG offers a low-cost, non-invasive alternative to the traditional electrocardiogram (ECG). Studies have demonstrated PPG's effectiveness in monitoring healthy individuals, suggesting its viability as a replacement for ECG in some settings (Bolanos et al. 2006 ). Leveraging embedded devices like Raspberry Pi and Arduino increases the functionality of PPG. Consistent with Borthakur (2020), we used the Pulse Sensor Amped, an Arduino-based heart-rate sensor, to collect PPG data from participants, recording from the initial baseline to the final baseline. Previous data, including analyses in both time and frequency domains, indicate that PPG can be as reliable as ECG when a sampling frequency of at least 25 Hz is used (Choi and Shin 2017 ). To optimize accuracy, participants were asked to minimize movement of the hand equipped with the sensor. We used the Python toolbox NeuroKit2 (Makowski, 2021) for data analysis. The raw PPG signal, sampled at 160.414 Hz, underwent preprocessing with a bandpass filter (0.5–8 Hz, Butterworth 3rd order, following Elgendi et al., 2013 ). For more accurate RR interval estimation from PPG signals, we deployed specific functions from NeuroKit2 based on methodologies reported by Orphanidou et al. ( 2015 ) and Frasch, M. G. ( 2022 ). Data Analysis Accept/Reject Decisions Behavioral data were processed using R Studio Version 2024.4.2.764. We defined rejection rates as the ratio of rejected offers to total offers in each category (Disadvantageous, Advantageous, Fair, and Overall). Offer types were categorized as Disadvantageous if the participant was offered 0–40%, Fair if offered 50%, and Advantageous if offered 60–100%. Given non-normality of the distribution of rejection rates, a Kruskal-Walli’s test was conducted to compare rejection rates between different offer types. The Wilcoxon Signed-Rank Test was used to assess differences in rejection rates between the Human and AI conditions for advantageous and disadvantageous offers. Consistent with previous approaches in the literature (e.g., Osumi & Ohira ( 2009 ), repeated measures ANOVAs were performed on the affect questionnaire data to evaluate the effect of offer type (Fair, Advantageous, Disadvantageous) on subjective feelings. Additionally, Spearman’s rank correlation was used to examine relationships between self-reported emotions, heart rate variability (HRV) metrics, and rejection rates and counts. To ensure consistency across affective measures, negative affects (Anger, Aversion, and Disgust) were reverse coded on a standardised scale, so that higher values indicated higher positive affect. Although ‘anger,’ ‘aversion,’ ‘reassurance,’ and ‘pleasure’ suggest four semantically distinct emotion concepts, analyses revealed that participants’ responses were highly intercorrelated: Cronbach’s 𝛂 = 0.86. Given this high degree of intercorrelation, we averaged the scores for each participant to create a composite affect index, with higher values indicating higher affect positivity. Physiological data analysis From the PPG recordings, we extracted the following time, frequency, and nonlinear measures using the neurokit2 toolbox: Standard Deviation of NN Intervals (SDNN), which indicates overall heart rate variability. Low Frequency (LF) and High Frequency (HF) components, which reflect a combination of sympathetic and parasympathetic activity, and parasympathetic activity, respectively. Cardiac Vagal Index (CVI), which specifically measures parasympathetic function. MeanNN represents the mean of all normal-to-normal (NN) intervals in milliseconds, which is inversely related to heart rate. Root Mean Square of the Successive Differences (RMSSD), which captures short-term variability linked to parasympathetic activity. pNN50, which is the percentage of NN intervals differing by more than 50 ms. Shannon Entropy, which measures the complexity and unpredictability of heart rate patterns. SD1 and SD2, which reflect short-term and long-term autonomic regulation, respectively, and are considered to be sensitive to dynamic changes in stress and recovery. We calculated these measures during the baseline period (5 min), then estimated it over three five-minute recording windows during the Ultimatum game play. The first and last 30 seconds of the data were excluded from the analysis to avoid potential artifacts or signal instability commonly observed during the start and end of PPG/ECG recordings, ensuring more reliable and accurate measurements. For short-term HRV measurements, 5-minute segments are generally considered appropriate, along with shorter durations (Cammet al. 1996). Results 1. Behavioral results: Accept/Reject Decisions 1.1 Rejection Rates Varied by Offer Type First, to ascertain that the data replicated previous findings (e.g., Osumi and Ohira ( 2009 ); Haselhuhn and Mellers, 2005 ; van't Wout et al., 2006 ), we compared the rejection rates by offer type (Advantageous, Fair, Disadvantageous), collapsing across the human and AI counterpart conditions. A Kruskal-Walli’s test revealed a significant difference in rejection rates among the different offer types in the predicted direction (i.e. highest for Disadvantageous, lowest for Advantageous), χ² (2) = 257, p < .001. 1.2 The Rejection Rate for Disadvantageous Offers Was Higher in the AI Condition Than in the Human Condition Next, we compared the rejection rates for disadvantageous offers between the human and AI conditions. A Wilcoxon rank-sum test revealed that participants were more likely to reject disadvantageous offers from an AI counterpart than from a human counterpart, W = 1688, p = .0067 (Fig. 2 A). Note that this is consistent with our hypothesis; whereas Sanfey et al. (2003) reported a higher rejection rate with the human counterpart, these data - collected over two decades later and with a much larger sample - indicate a reversal of the original effect. 1.3 The Advantageous Offer Rejection Rate Was Higher in the Human Condition: Next, we compared the rejection rates for advantageous offers between the human and AI conditions. A Wilcoxon rank-sum test indicated that participants exhibited a significantly higher rejection rate for advantageous offers when playing with a human counterpart, W = 2703.5, p = .007 (Fig. 2 B). This is also consistent with our hypothesis. 2. Impact of Affect 1.4 Subjective Affect Varied by Human vs AI Condition After removing mindless or logically incoherent entries (for more information on exclusion criteria, see Supplementary Material 4 ), we examined the effect of counterpart type and offer type on self-reported affect following fair offers, disadvantageous offers, and advantageous offers. The results are summarised in Figs. 3 , 4 and 5. Recall that higher values indicate higher affect positivity. ​​Disadvantageous Offers A one-way analysis of variance (ANOVA) on the self-reported affect index following disadvantageous offers revealed a significant effect of Counterpart Type (Human vs. AI), F (1, 128) = 6.79, p = 0.01, ηp² = 0.05. The direction of the effect indicates that participants reported feeling less positive affect / more negative affect following disadvantageous offers made by AI counterparts than by human counterparts. Advantageous Offers In contrast, an analogous ANOVA on self-reported affect following advantageous offers revealed no significant effect of Counterpart Type, F (1, 130) = 0.707, p = 0.402, ηp² = 0.005. This suggested no substantial difference in self-reported affect when participants received advantageous offers from human versus AI counterparts. Fair Offers: An analogous ANOVA for fair offers revealed no significant main effect of Counterpart Type (Human vs. AI) on self-reported affect, F (1, 130) = 1.981, p = 0.162, ηp² = 0.015. This indicates that participants’ emotional responses to fair offers did not differ significantly when the offers were made by a human counterpart or an AI counterpart. These results suggest that AI counterparts only elicit more negative affect than human counterparts following a disadvantageous offer. This pattern is consistent with the finding that participants were more likely to reject disadvantageous offers from the AI counterpart than from the human counterpart. 1.5 Subjective Affect Was Generally Correlated with Rejection Rate, But Did Not Vary as a Function of Counterpart Type (AI versus Human). To assess the relationship between subjective affect and rejection rate, we conducted correlation analyses between affective responses and rejection rates for each offer type, as depicted in Table 1 . Disadvantageous Offers: Following disadvantageous offers, affect positivity was negatively correlated with rejection rate in both the AI condition (rs = − 0.321, p < 0.01) and the Human condition (rs = − 0.497, p < 0.001). To investigate whether the relation between affect and rejection rate was stronger in the Human condition than in the AI condition, we regressed rejection rate onto affect positivity, counterpart type and the interaction term. Results indicated that affect positivity significantly predicted rejection rate, B = -0.089, SE = 0.022, p < .001, with the negative sign indicating that more positive affect was associated with lower rejection rate. However, the interaction between counterpart type and affect was not significant, B = 0.030, SE = 0.030, p = .325, suggesting that the relationship between affect and rejection rate did not differ as a function of playing with an AI or human counterpart. Advantageous Offers : For advantageous offers, affect positivity was negatively correlated with rejection in the AI condition (rs = − 0.291, p < 0.05) and in the Human condition (rs = − 0.465, p < 0.001). As above, we performed regression analysis to investigate whether participants experienced stronger emotional responses when rejecting advantageous offers in the Human condition compared to the AI condition. Results indicated a significant main effect of counterpart type, B = -0.071, SE = 0.028, p = .013, suggesting that participants rejected advantageous offers from human counterparts more frequently than from AI counterparts. Affect also significantly predicted rejection rate, B = -0.084, SE = 0.018, p < .001, with higher positive affect associated with lower rejection rates. The interaction between counterpart type and affect approached significance, B = 0.042, SE = 0.024, p = .078, indicating that, if anything, the effect of affect on rejection rate was stronger in the human condition, although this association should be interpreted with caution. Fair Offers : Affect positivity following fair offers was negatively correlated with rejection rate of fair offers in the AI condition, (rs = − 0.339, p < 0.01) and the Human condition (rs = − 0.477, p < 0.001). That is, the more negative (less positive) affect participants reported feeling after an offer, the more likely they were to reject that offer. To test for possible differences between the AI condition and the human condition, we conducted a regression analysis analogous to the ones above. It revealed that the main effect of counterpart type was not significant, B = -0.046, SE = 0.035, p = .194, indicating no significant difference in Fair offer rejection rates between AI and Human counterparts. However, affect significantly predicted rejection rate, B = -0.075, SE = 0.014, p < .001, with more positive affect associated with lower rejection rates. The interaction between counterpart type and affect was not significant, B = 0.027, SE = 0.024, p = .254, suggesting that the effect of affect on rejection rate did not differ between AI and Human counterpart. Table 1 Correlations between affect and offer rejection rates in the human and AI conditions. Significance levels: *p < .05, **p < .01, ***p < .001. Counterpart Type Fair RR Advantageous RR Disadvantageous RR AI -0.339** -0.291* -0.321** H -0.477*** -0.465*** -0.497*** In summary, analyses of participants’ self-reported affect revealed that participants reported feeling (a) generally more displeasure following disadvantageous offers than following the other types of offers and (b) more displeasure following disadvantageous offers from an AI counterpart than from a human counterpart. However, these differences did not differentially influence participants’ ultimate decisions to accept or reject offers from AI versus human counterparts. This may indicate that self-reported affective state is a comparatively crude measure, susceptible to both reporting biases and high variability. It may also indicate that additional processes, beyond affect play an important role in accept/reject decisions. We turn next to such potential processes. 2. Heart Rate Variability 2.1 Manipulation Check: Analysis of HRV Responses Across Initial Baseline, Ultimatum Game, and Final Baseline Periods First, to examine the effect of simply playing the Ultimatum Game on Heart Rate Variability (HRV), we analyzed HRV across three time periods: the initial baseline (B1), the UG period, and the final baseline (B2). Significant changes in HRV were observed across these periods, reflecting physiological variations. Because no significant interaction or main effect of counterpart type (Human vs. AI) was detected, we combined the conditions (H/AI) and conducted a repeated measures one way ANOVA to focus on the overall effect of time. For analysis, for consistency we focused on the middle five minutes of the UG condition as baseline and UG was of different length. Data from 12 participants were excluded due to missing the final baseline or both the ultimatum game (UG) and the final baseline caused by sensor data collection errors. A one-way ANOVA revealed a significant main effect of time for CVI, F (1.74, 190.09) = 17.76, p < .001, partial η² = .14. Pairwise comparisons demonstrated that CVI was significantly lower during the initial baseline compared to the second baseline, t (109) = − 5.08, p < .001, Cohen’s d = − 0.48. The initial baseline was not significantly different from the UG period, t (109) = − 2.35, p = .062, Cohen’s d = − 0.22. Additionally, CVI was significantly higher during the second baseline compared to the UG period, t (109) = 4.34, p < .001, Cohen’s d = 0.41. For RMSSD, a significant main effect of time was observed, F (1.66, 181.02) = 4.30, p = .015, partial η² = .038. Pairwise comparisons showed no significant difference between the initial and second baselines, t (109) = − 2.33, p = .066, Cohen’s d = − 0.22, nor between the initial baseline and the UG period, t (109) = − 0.71, p = .481, Cohen’s d = − 0.07. However, RMSSD was significantly higher in the second baseline compared to the UG period, t (109) = 2.53, p = .039, Cohen’s d = 0.24. For SD1, a significant main effect of time was observed, F (1.66, 181.03) = 4.31, p = .015, partial η² = .038. Pairwise comparisons indicated no significant difference between the initial baseline and the second baseline, t (109) = − 2.33, p = .065, Cohen’s d = − 0.22, or between the initial baseline and the UG period, t (109) = − 0.71, p = .481, Cohen’s d = − 0.07. The second baseline was significantly higher than the UG period, t (109) = 2.53, p = .039, Cohen’s d = 0.24. For SD2, a significant main effect of time was observed, F (1.83, 199.98) = 29.17, p < .001, partial η² = .211. Pairwise comparisons revealed that SD2 was significantly lower during the initial baseline compared to the second baseline, t (109) = − 6.57, p < .001, Cohen’s d = − 0.63. The initial baseline was significantly lower than the UG period, t (109) = − 2.55, p = .037, Cohen’s d = − 0.24. Additionally, SD2 was significantly higher during the second baseline compared to the UG period, t (109) = 5.34, p < .001, Cohen’s d = 0.51. For SDNN, a significant main effect of time was observed, F (1.78, 194.42) = 16.49, p < .001, partial η² = .131. Pairwise comparisons indicated that SDNN was significantly lower during the initial baseline compared to the second baseline, t (109) = − 5.44, p < .001, Cohen’s d = − 0.52. The initial baseline was not significantly different from the UG period, t (109) = − 2.01, p = .142, Cohen’s d = − 0.19. SDNN was significantly higher during the second baseline compared to the UG period, t(109) = 4.87, p < .001, Cohen’s d = 0.46. For Shannon Entropy, a significant main effect of time was observed, F (1.78, 194.42) = 21.25, p < .001, partial η² = .163. Pairwise comparisons revealed that ShanEn was significantly lower during the initial baseline compared to the second baseline, t (109) = − 4.85, p < .001, Cohen’s d = − 0.46. The difference between the initial baseline and the UG period was not significant, t (109) = − 1.97, p = .154, Cohen’s d = − 0.19. However, ShanEn was significantly higher in the second baseline compared to the UG period, t (109) = 4.24, p < .001, Cohen’s d = 0.40. Lower HRV during the initial baseline aligns with anticipatory stress before engaging in a task (Pulopulos et al., 2018 ). The UG period likely imposed cognitive or emotional challenges, reflected in reduced HRV measures such as SDNN, CVI, RMSSD, SD1, SD2. Similarly, Entropy reductions during stress periods (initial baseline and UG) likely reflect lower complexity in heart rate variability, a marker of reduced autonomic adaptability under stress (Pincus & Goldberger, 1994 ). Increased variability during the second baseline indicates the individual's recovery and adaptive capacity following the stressor, as suggested by research on autonomic recovery dynamics (Kim et al., 2018 ). These analyses provide evidence that HRV measures do capture heightened cognitive or emotional demands during the UG period. These data also serve as a manipulation check. The initial baseline's lower HRV likely reflects anticipatory stress, while the final baseline's higher HRV indicates recovery and relaxation following task completion. 2.2 Association between Heart Rate Variability and Ultimatum Game Decisions Next, we examined the relationship between various HRV measures and fair, advantageous, disadvantageous, and overall offer rejection rate in the Human and AI conditions. Spearman correlations are presented in Table 2 , as they are generally robust to outliers (Croux & Öllerer, 2015). The outliers were not removed, but rather winsorized following ​​the recommendations of Pulopulos, Vanderhasselt, and De Raedt ( 2018 ). The results reveal distinct patterns across various HRV metrics. In line with Sütterlin et al. ( 2011a ), who reported Pearson correlations, we also tested Pearson correlations. Detailed results, largely align with the Spearman correlation findings. At the aggregate level, a correlational analysis suggests differences in HRV metrics between AI and Human conditions. For overall rejection rate, correlations between HRV measures and rejection rate were consistently stronger in the AI condition: SDNN (0.334 for AI vs. 0.062 for H), SD2 (0.401 for AI vs. 0.098 for H), and ShanEn (0.312 for AI vs. 0.035 for H). These differences indicate that participants in the AI condition generally displayed a stronger association between physiological metrics and rejection of offers. Disadvantageous Offers For disadvantageous offers, a similar trend emerged, with stronger correlations in the AI condition for SD2 (0.201 for AI vs. -0.08 for H) and ShanEn (0.153 for AI vs. -0.114 for H). The negative correlations in the Human condition (e.g., MeanNN: -0.167 for H vs. 0.181 for AI) suggest differing physiological responses to unfairness when interacting with AI versus humans. To examine more directly whether the type of counterpart (human vs. AI) influenced the relationship between heart rate variability (HRV) predictors and rejection of disadvantageous offers, we conducted multiple regression analyses. We found no correlation between body mass index (BMI) and HRV, so we excluded it as a confounding variable in the subsequent regression analysis. We regressed the disadvantageous offer rejection rate onto each specific HRV measure, condition (Human vs. AI), and the interaction term. Significant interaction effects were observed for SDNN (β = -0.0028, p = .021), MeanNN (β = -0.0011, p = .002), SD2 (β = -0.0023, p = .040), and ShanEn (β = -0.1652, p = .003), such that participants in the AI condition exhibited a stronger positive association between these HRV metrics and rejection rate than did participants in the human condition. Advantageous Offers For advantageous offers, the differences between AI and Human conditions were less pronounced but still notable. Measures such as SD2 (0.161 for AI vs. 0.102 for H) and MeanNN (0.185 for AI vs. 0.212 for H) exhibited comparable correlations in both conditions, reflecting similar physiological responses to favorable outcomes. To examine more directly whether the type of counterpart (Human vs. AI) influenced the relationship between heart rate variability (HRV) measures and rejection of advantageous offers, we conducted analogous regression analyses. Rejection rate was regressed onto each HRV measure, Counterpart Type (Human vs. AI), and the interaction term. No significant interaction effects were observed for SDNN (β = -0.00005, p = .940), MeanNN (β = -0.00001, p = .947), SD2 (β = -0.00009, p = .892), or ShanEn (β = -0.0011, p = .974), indicating that the associations between these HRV measures and rejection rates did not differ between the Human and AI conditions. A significant main effect of MeanNN was observed, β = 0.0004, p = .006, suggesting a positive association between MeanNN and rejection rate regardless of whether the counterpart was a human or an AI agent. Additionally, counterpart type significantly predicted rejection rates, β = -0.052, p = .005, with participants in the Human condition rejecting advantageous offers more frequently than participants in the AI condition. Neither Affect (β = -0.002, p = .813) nor its interaction with SDNN (β = -0.0002, p = .327) were significant predictors of rejection rate. Fair Offers For fair offers, the Spearman correlation analysis revealed stronger associations in the AI condition compared to the Human condition for most HRV measures. Notable differences include SDNN (0.377 for AI vs. 0.068 for H), SD2 (0.405 for AI vs. 0.099 for H), and ShanEn (0.344 for AI vs. 0.077 for H). To examine more directly whether the type of counterpart (Human vs. AI) influenced the relationship between heart rate variability (HRV) measures and rejection of fair offers, we conducted analogous multiple regression analyses. Rejection rate was regressed onto each HRV measure, Counterpart Types (Human vs. AI), and their interaction term. Significant interaction effects were observed for SDNN (β = -0.0023, p = .016) and SD2 (β = -0.0021, p = .014), indicating that these HRV measures were more strongly associated with rejection rate in the AI condition compared to the human condition. Table 2 "Comparison of Spearman correlations between HRV measures (SDNN, SD1, SD2, RMSSD, MeanNN, ShanEn, and CVI) and Ultimatum Game decisions (rejection rate for Fair offers, Disadvantageous offers, and Overall Rejection Rate) in the Human and AI conditions. Significant differences highlight stronger associations in the AI condition across multiple HRV metrics. Measure Fair (H) Fair (AI) Disadv (H) Disadv (AI) Adv (H) Adv (AI) Overall RR (H / AI) SDNN 0.068 0.377*** -0.106 0.160* 0.122 0.138 0.062 / 0.334*** SD1 0.005 0.338*** -0.074 0.145 0.069 0.069 0.056 / 0.264*** SD2 0.099 0.405*** -0.080 0.201** 0.102 0.161* 0.098 / 0.401*** RMSSD 0.005 0.338*** -0.074 0.145 0.069 0.069 0.056 / 0.264*** MeanNN -0.074 0.077 -0.167* 0.181* 0.212* 0.185* -0.031/ 0.257*** ShanEn 0.077 0.344*** -0.114 0.153* 0.095 0.096 0.035 / 0.312*** CVI 0.022 0.340*** -0.088 0.157* 0.120 0.107 0.054 / 0.294*** In summary, these analyses reveal stronger HRV associations with rejection rate in the AI condition than in the Human condition, suggesting that individuals experience heightened physiological activation when interacting with an AI agent (Fig. 7 ). The inhibitory component of rejecting a disadvantageous offer is thought to involve overriding one’s default preference to accept any money that is offered (Sutterlin et al., 2011). Thus, these analyses suggest that participants’ higher rejection rate of disadvantageous offers from the AI counterpart is driven, in part, by higher parasympathetic system activity associated with inhibitory control. Was this association moderated by participants’ affect? If so, it would suggest that emotional tone contributes to shaping participants’ deployment of self-regulatory processes. If not, it would suggest that affective processes and self-regulatory processes represent largely independent contributions to accept/reject decisions. We turn next to this question. In regression analyses of rejection rates of Advantageous, Disadvantageous, and Fair offers with Affect, HRV (SDNN), and the interaction term as predictors, distinct patterns emerged between participants who faced human versus AI counterparts. For Advantageous offers, participants in the human condition showed no significant main effects for SDNN (β = 0.00078, p = 0.220) or Affect (β = -0.00476, p = 0.541), but a significant interaction between SDNN and Affect (β = -0.00060, p = 0.034), suggesting that as parasympathetic activity (SDNN) increases, the relationship between Affect and rejection rate weakens or reverses. In contrast, participants in the AI condition displayed no significant main effects of SDNN (β = 0.00060, p = 0.156) or Affect (β = 0.00102, p = 0.843), and no interaction effects (β = 0.00011, p = 0.623). For Disadvantageous offers, participants in the human condition exhibited a significant main effect of Affect (β = -0.0266, p = 0.023), indicating that greater negative affect predicted higher rejection rates, while SDNN (β = -0.00057, p = 0.549) and the interaction term (β = -0.00005, p = 0.898) were non-significant. In the AI condition, in contrast, SDNN showed a significant positive main effect (β = 0.00174, p = 0.033), suggesting that higher parasympathetic activity was associated with increased inhibitory control when rejecting disadvantageous offers, whereas Affect (β = -0.0115, p = 0.252) and the interaction term (β = 0.00018, p = 0.675) were not significant. For Fair offers, participants in the human condition displayed a significant main effect of Affect (β = -0.0286, p = 0.008), indicating that greater negative affect was associated with higher rejection rates, while SDNN (β = 0.0012, p = 0.159) and the interaction (β = -0.00050, p = 0.184) were not significant. In the AI condition, SDNN emerged as a significant predictor of rejection rate (β = 0.00279, p < 0.001), while Affect (β = -0.0064, p = 0.310) and the interaction term (β = -0.00025, p = 0.365) were non-significant. In summary, these analyses suggest distinct mechanisms driving rejection decisions when interacting with human versus AI counterparts. Specifically, the decision to reject an offer appears to be influenced by two sources: (a) affective experience and (b) inhibition of the default tendency to accept any amount of money. With a human counterpart, the decision to reject an offer is primarily driven by affective state, such that higher positive emotion reduces the likelihood of rejection. This reflects a relatively straightforward process, as participants can readily interpret the experience of receiving a disadvantageous offer within a typical, human-to-human, normative framework. In contrast, when interacting with an AI counterpart, rejection decisions are less influenced by emotional states, but more influenced by the deployment of inhibitory processes—reflected in higher parasympathetic activity (e.g. SDNN)—which override the default tendency to accept any offer. This additional inhibitory control may stem from participants' unfamiliarity and uncertainty with how to navigate social conventions with a quasi-human entity, making the act of rejecting an AI offer comparatively more effortful. Figure 8 illustrates this proposed hydraulic relation between affect and self-regulation by enlarging the affect component for human counterparts, but enlarging the self-regulation component for AI counterparts. General Discussion These data yielded several novel findings regarding individuals’ behavioral, physiological, and affective responses to disadvantageous, advantageous, and fair offers from AI agents. First, participants were more likely to reject disadvantageous offers from an AI counterpart than from a human counterpart. This pattern directly contrasts with findings reported by Sanfey et al. (2003) (and was obtained with a much larger sample). It is attributable, we suggest, to at least four reasons. First, with the passage of time, AI entities have assumed a larger role in millions of people’s everyday lives. This has likely led to evolving expectations of AI social behavior since 2003. Second, research on the Perfect Automation Schema (Merrit et al., 2015) has demonstrated that people often expect higher reliability and impartiality from autonomous agents. For example, Shariff, Bonnefon, and Rahwan ( 2021 ) reported that participants held autonomous vehicles to a higher standard of safety than they required from human drivers. Third, people may believe that AI agents should be subservient to humans (De Freitas et al, 2023 ); thus, an AI agent that attempts to domineer a human with a lowball offer would violate its presumed subservient position. These findings align with Treiman et al. ( 2023 ), who demonstrated that participants even exhibit a willingness to incur personal costs to instill fairness in AI. Fourth, although LLMs are increasingly convincing as conversational agents, the social norms that inhibit negative emotional expressions in human-to-human interactions may nonetheless be weaker in human-to-AI interactions (Oudah et al., 2024 ). Thus, people may feel freer to express displeasure toward - or even ‘punish’ - an entity that will presumably not take offence. Influence of HRV versus Affect Although participants generally reported feeling worse after receiving disadvantageous offers, self-reported positive/negative affect was equally associated with the rejection of disadvantageous offers from human and AI counterparts. In contrast, the relation between heart rate variability (HRV) metrics such as MeanNN, SDNN, and RMSSD and rejection rate displayed a clear sensitivity to counterpart type (AI vs. Human), with stronger associations between parasympathetic engagement and rejection of offers from AI counterparts than from human counterparts. These HRV metrics are commonly thought to index individuals’ level of effortful focus. Thus, the experience of playing the Ultimatum Game with an AI agent, and receiving a lowball offer from an AI agent, appear to intensify self-control processes. Taken together, we propose the following explanatory model (depicted in Fig. 8 ): In typical, human-human interactions, societal fairness norms act as a check against the default tendency to accept any money that is offered. Such culturally ingrained fairness norms contribute to the rejection of both disadvantageous offers (“You are ripping me off!”) and advantageous offers (“I am ripping you off!”). In human-AI interactions, however, such social norms have yet to be clearly established and are thus weaker. In such cases, to restrain the default inclination to accept a disadvantageous offer, additional input is needed. This input appears to come from effortful, self-regulatory processes, captured by measures of HRV. In other words, in the absence of traditional emotional cues, physiological regulation reflected in HRV may serve as an important complementary mechanism that contributes to the decision to accept or reject the offer. This model aligns with prior research that has emphasized the role of HRV in regulating fairness-related behavior (Sütterlin et al., 2011). Advantageous Inequity Aversion (AIA): Advantageous offers, despite being beneficial, are perceived as violations of fairness norms, leading to discomfort linked to guilt or embarrassment. Participants in this study demonstrated higher rejection rates of advantageous offers from human counterparts than from AI counterparts. This finding resonates with work by Shaw and Choshen-Hillel ( 2017 ) and McAuliffe et al. ( 2013 ), who provided evidence for the role of societal norms and social conditioning in shaping fairness perceptions. One possible contributor to this phenomenon is emotional synchrony. This phenomenon of emotional synchrony was captured by Adam Smith in his Theory of Moral Sentiments , in which he describes how individuals moderate their emotional expressions (including positive emotions) to align with societal expectations: “To see the emotions of their hearts, in every respect, beat in time to his own constitute his sole consolation. But he can only hope to obtain this by lowering his passion to that pitch in which the spectators are capable of going along with him. He must flatten, if I may be allowed to say so, the sharpness of its natural tone…” [Smith, 1759 /2009, p. 28]. That is, people often “flatten” their positive emotional tone when interacting with other humans, to avoid eliciting resentment or discomfort. This emotional calibration, however, becomes less relevant in interactions with AI agents, who are perceived, as the default, to be generally lower in - or devoid of - emotional capacity (Ayad & Plaks, 2025; Yam et al., 2020 ). Although the current procedure did not include any measures to directly test this positive emotion suppression mechanism, we consider it to be a plausible partial explanation for the diminished rejection rates of advantageous offers and the lower emotional disturbance when dealing with AI agents. We encourage future researchers to test this possibility more directly. Our findings extend prior research by demonstrating the critical role of inhibitory control and emotional regulation, measured by HRV, in navigating social norms of fairness with AI agents. Previous studies (e.g., Sanfey et al., 2003; Knoch et al., 2006) highlighted the importance of executive functioning in regulating fairness-related behavior, often mediated by the dorsolateral prefrontal cortex. The present study suggests that such self-regulatory processes captured by HRV compensate for the absence of external social cues and well-established social norms in AI interactions. In short, whereas emotional responses play the dominant role in human-human UG exchanges, self-regulation appears to assume a larger role when the counterpart is an AI agent. We suggest that in future studies, HRV measures might emerge as effective indices for capturing decision-making in human-AI interactions, particularly in contexts involving fairness. The findings have significant implications for the design and implementation of AI systems in social and economic contexts. Understanding how people perceive fairness in AI interactions can guide the development of systems that align with human moral and ethical standards. For example, embedding human-like social cues in AI agents could mitigate the reliance on physiological regulation, fostering trust and cooperation (Plaks, Bustos-Rodriguez, & Ayad, 2022). Moreover, as AI becomes more integrated into daily life, shifting cultural norms surrounding human-AI interaction must be continually assessed. Finally, clear cultural differences in comfort with robotic and AI agents have been documented (Castelo & Sarvary, 2022 , Lim et al., 2021 ). Future researchers will need to be mindful of participants’ cultural context when assessing their perceptions of AI fairness. Declarations ACKNOWLEDGMENTS: The authors acknowledge the valuable insights provided by members of the University of Toronto Motivation and Social Cognition laboratory. This work is partially funded by grants from the Natural Sciences and Engineering Research Council of Canada (fund # 508977) and the Schwartz-Reisman Institute for Technology and Society (SRI- VPRI). References Alnaeb, M. E., Alobaid, N., Seifalian, A. M., Mikhailidis, D. P., and Hamilton, G. (2007). Optical techniques in the assessment of peripheral arterial disease. Current vascular pharmacology, 5(1), 53–59. Ayad, R. & Plaks, J.E. (2025). Attribution of intent and moral responsibility to AI agents. Computers in Human Behavior: Artificial Humans, 3, 100107. Axelrod, R. (2006). The Evolution of Cooperation (Revised ed.). Basic Books. Beckers, F., Verheyden, B., & Aubert, A. E. (2006). Aging and nonlinear heart rate control in a healthy population. American Journal of Physiology-Heart and Circulatory Physiology, 290(6), H2560–70. doi:10.1152/ajpheart.00903.2005 Bishara, A. J., & Hittner, J. B. (2012). Testing the significance of a correlation with nonnormal data: comparison of Pearson, Spearman, transformation, and resampling approaches. Psychological methods, 17(3), 399. Blake, P. R., & McAuliffe, K. (2011). “I had so much it didn’t seem fair”: Eight-year-olds reject two forms of inequity. Cognition, 120(2), 215-224. Bohnet, I., & Zeckhauser, R. (2004). Trust, risk and betrayal. Journal of Economic Behavior & Organization, 55(4), 467–484. https://doi.org/10.1016/j.jebo.2003.11.004 Bolanos, M., Nazeran, H., & Haltiwanger, E. (2006). Comparison of heart rate variability signal features derived from electrocardiography and photoplethysmography in healthy individuals. 2006 International Conference of the IEEE Engineering in Medicine and Biology Society, IEEE, 4289–4294. Bolton, G. E., & Ockenfels, A. (2000). ERC: A theory of equity, reciprocity, and competition. American Economic Review , 90(1), 166–193. Castelo, N., & Sarvary, M. (2022). Cross-cultural differences in comfort with humanlike robots: Evidence from Japan and the United States. International Journal of Social Robotics. Chen, D. L., Schonger, M., & Wickens, C. (2016). oTree—An open-source platform for laboratory, online, and field experiments. Journal of Behavioral and Experimental Finance, 9(C), 88–97. Choi, A., & Shin, H. (2017). Photoplethysmography sampling frequency: Pilot assessment of how low can we go to analyze pulse rate variability with reliability? Physiological measurement, 38(3), 586. Christophe, C., & Öllerer, V. (2015). Robust and sparse estimation of the inverse covariance matrix using rank correlation measures. Social Science Research Network, 35-55. doi:10.1007/978-81-322-3643-6_3 Croux, C., & Öllerer, V. (2016). Robust and sparse estimation of the inverse covariance matrix using rank correlation measures (pp. 35-55). Springer India. De Freitas, J., Agarwal, S., Schmitt, B., & Haslam, N. (2023). Psychological factors underlying attitudes toward AI tools. Nature Human Behaviour, 7(11), 1845-1854. Dunning, D., Anderson, J. E., Schlösser, T., Ehlebracht, D., & Fetchenhauer, D. (2014). Trust at zero acquaintance: more a matter of respect than expectation of reward. Journal of Personality and Social Psychology, 107(1), 122. Dulleck, U., Schaffner, M., & Torgler, B. (2011). Heartbeat and economic decisions: Observing mental stress among proposers and responders in the ultimatum bargaining game. PLOS ONE. Elgendi, M., Norton, I., Brearley, M., Abbott, D., & Schuurmans, D. (2013). Systolic peak detection in acceleration photoplethysmograms measured from emergency responders in tropical conditions. PloS one, 8(10), e76585. Evans, A. M., & Revelle, W. (2008). Survey and behavioral measurements of interpersonal trust. Journal of Research in Personality , 42(6), 1585–1593. Evans, M., & Geddes, L. (1988). An assessment of blood vessel vasoactivity using photoplethysmography. Medical instrumentation, 22(1), 29–32. Frasch, M. G. (2022). Comprehensive HRV estimation pipeline in Python using Neurokit2: Application to sleep physiology. MethodsX, 9, 101782. Goldberger, A. L. (1991). Is the normal heartbeat chaotic or homeostatic? News Physiol Sci, 6, 87–91. Güth, W., Schmittberger, R., & Schwarze, B. (1982). An experimental analysis of ultimatum bargaining. Journal of Economic Behavior and Organization , 3(4), 367–388. Harsanyi, J. C. (1961). On the rationality postulates underlying the theory of cooperative games. Journal of Conflict Resolution , 5(2), 179–196. Haselhuhn, M.P., Mellers, B.A., 2005. Emotions and cooperation in economic games. Brain Res. Cogn. Brain Res. 23, 24–33. Inbar, Y., Pizarro, D. A., & Bloom, P. (2009). Conservatives are more easily disgusted than liberals. Cognition and Emotion, 23(4), 714–725. Jean-François Bonnefon, Iyad Rahwan, and Azim Shariff, “The Moral Psychology of Artificial Intelligence”, Annual Review of Psychology, vol. 75, January 2024, pp. 653–675. Johnson-George, C., & Swap, W. C. (1982). Measurement of Specific Interpersonal Trust: Construction and Validation of a Scale to Assess Trust in a Specific Other . Journal of Personality and Social Psychology, 43(6), 1306–1317. Karpus, J., Krüger, A., Verba, J. T., Bahrami, B., & Deroy, O. (2021). Algorithm exploitation: Humans are keen to exploit benevolent AI. Iscience, 24(6). Kim, H. G., Cheon, E. J., Bai, D. S., Lee, Y. H., & Koo, B. H. (2018). Stress and heart rate variability: a meta-analysis and review of the literature. Psychiatry investigation, 15(3), 235. Kosfeld, M., Heinrichs, M., Zak, P. J., Fischbacher, U., & Fehr, E. (2005). Oxytocin increases trust in humans. Nature, 435(7042), 673–676. Lim, A., Louie, W.-Y. G., & Su, N. M. (2021). Social robots on a global stage: Establishing a role for culture during human–robot interaction. International Journal of Social Robotics, 13(4), 1023–1037. Makovi K, Sargsyan A, Li W, Bonnefon JF, Rahwan T. 2023. Trust within human-machine collectives depends on the perceived consensus about cooperative norms. Nature Communications Makowski, D., Pham, T., Lau, Z. J., Brammer, J. C., Lespinasse, F., Pham, H., ... & Chen, S. A. (2021). NeuroKit2: A Python toolbox for neurophysiological signal processing. Behavior research methods, 1-8. Malle, B. F., & Ullman, D. (2021). A multidimensional conception and measure of human-robot trust. In Trust in human-robot interaction (pp. 3-25). Academic Press. McAuliffe, K., Blake, P. R., Kim, G., Wrangham, R. W., & Warneken, F. (2013). Social influences on inequity aversion in children. PLOS ONE, 8(12), e80966. McAuliffe, K., Blake, P. R., & Warneken, F. (2014). Children reject inequity out of spite. Biology letters, 10(12), 20140743. McCraty, R., & Shaffer, F. (2015). Heart rate variability: new perspectives on physiological mechanisms, assessment of self-regulatory capacity, and health risk. Global Advances in Health and Medicine, 4, 46–61. doi:10.7453/gahmj.2014.073 Morewedge, C. K., Krishnamurti, T., Ariely, D. (2014). "Focused on fairness: Alcohol intoxication increases the costly rejection of inequitable rewards". Journal of Experimental Social Psychology. 50: 15–20. doi:10.1016/j.jesp.2013.08.006 Merritt, S. M., Unnerstall, J. L., Lee, D., & Huber, K. (2015). Measuring individual differences in the perfect automation schema. Human factors, 57(5), 740-753. Nielsen YA, Pfattheicher S, Keijsers M. 2022a. Prosocial behavior toward machines. Current Opinion in Psychology 43:260–265 Nowak, M. A. , Page, K.M., & Sigmund, K. (2000). "Fairness Versus Reason in the Ultimatum Game". Science. 289 (5485): 1773–1775. Nowak, M. A., & Sigmund, K. (2005). Evolution of indirect reciprocity. Nature , 437(7063), 1291–1298. Oliveira, F. T., McDonald, J. J., & Goodman, D. (2007). Performance monitoring in the anterior cingulate is not all error related: expectancy deviation and the representation of action-outcome associations. Journal of Cognitive Neuroscience, 19(12), 1994-2004. Orphanidou, C., Bonnici, T., Charlton, P., Clifton, D., Vallance, D., & Tarassenko, L. (2015). Signal-Quality Indices for the Electrocardiogram and Photoplethysmogram: Derivation and Applications to Wireless Monitoring . IEEE Journal of Biomedical and Health Informatics, 19(3), 832–838. Osumi, T., & Ohira, H. (2009). Cardiac responses predict decisions: An investigation of the relation between orienting response and decisions in the ultimatum game. Biological Psychology, 82(2), 174-181. Oudah, M., Makovi, K., & Gray, K. (2024). Perception of experience influences altruism and perception of agency influences trust in human–machine interactions. Scientific Reports , 14, 63360. Pettit, P. (2018). The Birth of Ethics: Reconstructing the Role and Nature of Morality . Oxford University Press. Pincus, S. M., & Goldberger, A. L. (1994). Physiological time-series analysis: What does regularity quantify? American Journal of Physiology-Regulatory, Integrative and Comparative Physiology, 266(4), R1395-R1407. Plaks, J. E., Rodriguez, L. B., & Ayad, R. (2022). Identifying psychological features of robots that encourage and discourage trust. Computers in Human Behavior, 134, 107301. Pulopulos, M. M., Vanderhasselt, M. A., & De Raedt, R. (2018). Association between changes in heart rate variability during the anticipation of a stressful situation and the stress-induced cortisol response. Psychoneuroendocrinology, 94, 63-71. Robinson, J.S., Xu, X. & Plaks, J.E. (2019). Disgust and deontology: Trait sensitivity to pathogens promotes a preference for clarity, hierarchy, and rule-based moral judgment. Social Psychological and Personality Science, 10 , 3-14 . Rosas, A., Bermúdez, J. P., Cotrina, J. M., Aguilar-Pardo, D., Caicedo, J. C., & Aponte-Canencio, D. M. (2020). Perceiving utilitarian gradients: Heart rate variability and self-regulatory effort in the moral dilemma task. Journal of Moral Education, 49(1), 45-61. Rotter, J. B. (1967). A new scale for the measurement of interpersonal trust. Journal of Personality , 35(4), 651–665. Santamaría-García, H., Cotrina, J. M., Torres, N. F., Buitrago, C., Aponte-Canencio, D. M., Caicedo, J. C., Billeke, P., Gantiva, C., & Baez, S. (2020). Explicit and implicit markers of fairness preeminence in criminal judges. Scientific Reports, 10, 11234. Schall, M., Martiny, S. E., Goetz, T., & Hall, N. C. (2016). Smiling on the inside: The social benefits of suppressing positive emotions in outperformance situations. Personality and Social Psychology Bulletin, 42, 559–571. Shaffer, F., & Ginsberg, J. P. (2017). An overview of heart rate variability metrics and norms. Frontiers in Public Health, 5, 258. Shariff, A., Bonnefon, J.J. & Rahwan, I. (2021). How safe is safe enough? Psychological mechanisms underlying extreme safety demands for self-driving cars, Transportation Research Part C: Emerging Technologies, 126, 103069. Shaw, A., & Choshen-Hillel, S. (2017). It’s not fair: Folk intuitions about disadvantageous and advantageous inequity aversion. Judgment and Decision Making, 12(3), 208-223. Sheldon, O., Plaks, J.E., Sridharan, V., & Shoda, Y. (2018). Strategic actors in situ impressions of systematically- versus unsystematically-variable counterparts. Social Cognition, 36, 324-344. Sloane, S., Baillargeon, R., & Premack, D. (2012). Do infants have a sense of fairness? Psychological Science, 23(2), 196–204. Smith, A.M., Young, G. & Ford, B.Q. (2023). The interpersonal correlates of believing emotions are controllable. Motivation and Emotion, 47, 323–332. Smith, A. (1759). The Theory of Moral Sentiments . Stein, P. K., & Reddy, A. (2005). Non-linear heart rate variability and risk stratification in cardiovascular disease. Indian Pacing Electrophysiology Journal, 5, 210–20. Strohminger, N., & Jordan, M. R. (2022). Corporate insecthood. Cognition, 224, 105068. Sütterlin, S., Herbert, C., Schmitt, M., Kübler, A., & Vögele, C. (2011a). Overcoming selfishness: reciprocity, inhibition, and cardiac-autonomic control in the ultimatum game. Frontiers in Psychology, 2, 11221. Sütterlin, S., Herbert, C., Schmitt, M., Kübler, A., & Vögele, C. (2011b). Frames, decisions, and cardiac–autonomic control. Social Neuroscience, 6(2), 169-177. Treiman, L. S., Ho, C. J., & Kool, W. (2023, November). Humans forgo reward to instill fairness into AI. In Proceedings of the AAAI Conference on Human Computation and Crowdsourcing (Vol. 11, No. 1, pp. 152-162). van 't Wout, M., Kahn, R. S., Sanfey, A. G., & Aleman, A. (2006). Affective state and decision-making in the Ultimatum Game. Experimental Brain Research, 169(4), 564-8. doi: 10.1007/s00221-006-0346-5. Walster, E., Walster, G. W., & Berscheid, E. (1978). Equity: Theory and research . Allyn & Bacon Waytz, A., Cacioppo, J., & Epley, N. (2010). Who Sees Human? The Stability and Importance of Individual Differences in Anthropomorphism . Perspectives on Psychological Science, 5(3), 219–232. Wu, Y., & Zhou, X. (2009). The P300 and reward valence, magnitude, and expectancy in outcome evaluation. Brain Research, 1286, 114-122. Yam, K. C., Bigman, Y. E., Tang, P. M., Ilies, R., De Cremer, D., Soh, H., & Gray, K. (2020). Robots at work: People prefer—and forgive—service robots with perceived feelings. Journal of Applied Psychology, 106(10), 1557–1572. Zou, C., Plaks, J.E., & Peterson, J.B. (2019). Don’t get too excited: Assessing individual differences in the downregulation of positive emotions. Journal of Personality Assessment, 101, 73-83. Footnotes Such games raise a reasonable question: What would an AI agent do with any money it earns? The data thus far, however, indicate that participants are generally untroubled by this question (e.g., Makovi et al., 2023 ; Plaks et al., 2022 ), likely assuming that although the money may be useless to the agent, it nonetheless has been programmed to attempt to maximize its earnings (Bonnefon et al., 2024).. The age data were available for 136 participants. One measure assessed trait-level dehumanization tendencies (Yam et al., 2021). A second measure assessed individual differences in anthropomorphism tendencies (Waytz et al., 2010 ). A third measure was a single item assessing the perceived importance of fairness in decision-making. A fourth measure assessed trait-level interpersonal trust Johnson-George & Swap, 1982 ). The results did not meaningfully vary whether the excluded values were included or excluded. Additional Declarations There is NO Competing Interest. Supplementary Files supplementarymaterial.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-6051145","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":424093441,"identity":"6e2d79a7-1ed3-4e9c-a269-0a8a9f3e9c13","order_by":0,"name":"Debanjan Borthakur","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0001-8651-7569","institution":"University of Toronto","correspondingAuthor":true,"prefix":"","firstName":"Debanjan","middleName":"","lastName":"Borthakur","suffix":""},{"id":424093442,"identity":"8c533fa6-3b0f-441d-98cd-d0df64c5a6a8","order_by":1,"name":"Peter Diep","email":"","orcid":"","institution":"University of waterloo","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"","lastName":"Diep","suffix":""},{"id":424093443,"identity":"0459d85d-05a8-4d4e-a5d8-aac3338756f2","order_by":2,"name":"Jason Plaks","email":"","orcid":"","institution":"University of Toronto","correspondingAuthor":false,"prefix":"","firstName":"Jason","middleName":"","lastName":"Plaks","suffix":""}],"badges":[],"createdAt":"2025-02-17 22:45:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6051145/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6051145/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":79259057,"identity":"a0386638-7aee-4156-ab7c-afe812563e67","added_by":"auto","created_at":"2025-03-26 09:16:59","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":321117,"visible":true,"origin":"","legend":"\u003cp\u003eExperimental protocol overview. Participants were randomized into Human versus AI conditions. Both conditions involved a 5-minute baseline PPG recording, 21 offers in the Ultimatum Game, a 5-minute final baseline, and questionnaires. Continuous PPG recording was maintained throughout.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6051145/v1/cf6f9193fa1e8f4a15f1156a.jpeg"},{"id":79256634,"identity":"6cc73165-dc97-434a-9e47-b066ff472fe1","added_by":"auto","created_at":"2025-03-26 09:00:59","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":39779,"visible":true,"origin":"","legend":"\u003cp\u003ePlot (A) shows the counterpart type comparison of the rejection rate for disadvantageous offers. Plot (B) shows the counterpart type comparison of rejection rate for advantageous offers. Significance levels: *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6051145/v1/8a690f6a3b9a59d7e2ba035a.png"},{"id":79256635,"identity":"91c552a3-0d18-4853-8bcc-b925a0d0876e","added_by":"auto","created_at":"2025-03-26 09:00:59","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":410932,"visible":true,"origin":"","legend":"\u003cp\u003eAffect by Counterpart Type. Higher values on the Y axis indicate higher positive affect. The figure summarizes the pairwise comparisons for (A) Disadvantageous offers. (B) Fair offers. (C) Advantageous offers Significance levels: *p \u0026lt; .05, **p \u0026lt; .01, ***p \u0026lt; .001.\u003c/p\u003e","description":"","filename":"floatimage314.png","url":"https://assets-eu.researchsquare.com/files/rs-6051145/v1/58e4be0a89da198e4afdbcc8.png"},{"id":79259055,"identity":"eaf6561b-6eca-414b-9136-475bcfbfd62d","added_by":"auto","created_at":"2025-03-26 09:16:59","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":645846,"visible":true,"origin":"","legend":"\u003cp\u003eFigure 7: Counterpart Type differences in the association between HRV indices and disadvantageous offer rejection rates.\u003cem\u003e \u003c/em\u003eThe interaction between Condition (Human = H, AI = Artificial Intelligence) and four heart rate variability (HRV) indices is shown: (A) SDNN (B) Shannon Entropy (ShanEn), (C) SD2 and (D) MeanNN. Blue solid lines represent the AI group, and orange dashed lines represent the Human group. Shaded regions indicate 95% confidence intervals. While HRV indices (ShanEn, SDNN, SD2, MeanNN) generally predict increasing rejection rates in the AI group, the Human group shows flatter or decreasing trends.\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6051145/v1/1c59032aee7bb7adece7d731.jpeg"},{"id":79259056,"identity":"b4c3214a-5e7c-4e0c-ba13-018cf4b55574","added_by":"auto","created_at":"2025-03-26 09:16:59","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":255568,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 8: \u003c/strong\u003eInteraction Model illustrating hypothesized decision-making processes in Human-Human and Human-AI UG decision making.\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6051145/v1/dcdc1cca759ec53f8388e2ef.jpeg"},{"id":79259061,"identity":"cbd7939d-dfc3-4145-9a1d-12344bb70516","added_by":"auto","created_at":"2025-03-26 09:17:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2564196,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6051145/v1/af9d67d0-6896-4eda-8151-c70a32e723f5.pdf"},{"id":79256639,"identity":"56fe8f72-7992-49cc-9456-a84edb6dbaf4","added_by":"auto","created_at":"2025-03-26 09:00:59","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":28858,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarymaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-6051145/v1/b495986dbe486680b499c4ef.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Inequity Aversion Toward AI Counterparts","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe implicit or explicit assumption of fair treatment is a cornerstone of human moral behavior (Pettit, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Sloane, Baillargeon, \u0026amp; Premack, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Numerous studies have provided evidence that individuals are more likely to behave prosocially to the extent that they trust that others will allocate resources equitably and will not systematically abuse such trust (e.g., Nowak \u0026amp; Sigmund, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Sheldon et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The research literatures on resource allocation and on interpersonal trust have, understandably, focused almost exclusively on human-to-human interactions (e.g., Axelrod, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Evans \u0026amp; Revelle, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Kosfeld et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Rotter, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e1967\u003c/span\u003e). However, with the increasing capabilities of LLM-based social agents to simulate human behavior, a growing literature has begun to turn these questions toward humans\u0026rsquo; moral expectations of AI agents with the aim of better understanding the mechanisms - and limits - of anthropomorphism (e.g., Karpus et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Nielsen et al., 2022b; Oudah et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Malle \u0026amp; Ullman, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Plaks, Bustos-Rodriguez \u0026amp; Ayad, 2022).\u003c/p\u003e\u003cp\u003eSeveral research teams have approached these questions using classic behavioral economic paradigms\u003csup\u003e1\u003c/sup\u003e. For example, Karpus et al. (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) reported that in a one-shot Prisoner\u0026rsquo;s Dilemma game, participants expected the same degree of cooperative behavior from an AI counterpart as from a human, although they themselves responded significantly more cooperatively toward the human (49%) than toward the AI counterpart (36%). (See also Plaks et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2022\u003c/span\u003e.) Similarly, in a one-shot Trust Game, participants expected equivalent payouts from a human or an AI counterpart, although they themselves reciprocated prosocially 75% of the time when their counterpart was human, but only 34% when it was an AI agent. Karpus et al. (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) reported analogous results with a one-shot Dictator Game and Nielsen et al. (2022b) reported parallel findings with a Public Goods game. In summary, although participants tend to approach an AI counterpart with the assumption that it will treat them as equitably as a human will, they feel less obligated to respond in kind toward the AI counterpart (Bonnefon, Rahwan, \u0026amp; Shariff, 2024).\u003c/p\u003e\u003cp\u003e\u003cb\u003eThe Ultimatum Game\u003c/b\u003e\u003c/p\u003e\u003cp\u003eAlthough the Prisoner\u0026rsquo;s Dilemma, Trust Game, and Dictator Game all partially concern expectations of fair treatment, they also invoke a range of further concerns, including interpersonal trust and strategic gamesmanship. Arguably a more direct method to assess assumptions of fair treatment \u003cem\u003eper se\u003c/em\u003e is the Ultimatum Game (UG) (G\u0026uuml;th et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e1982\u003c/span\u003e; Harsanyi, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1961\u003c/span\u003e). In a one-shot UG, the Proposer starts with a sum of money. The Proposer decides how to split the money with the Responder (who is made aware of the total sum of money). The Responder may accept or reject this split. If the Responder accepts it, the money is split accordingly. If the Responder rejects the split, both players receive nothing. Before play begins, both players are informed of all possible outcomes. Thus, if the Responder rejects a disadvantageous split (e.g., 20%-80%) - thereby choosing to receive nothing rather than something - this can be interpreted as a signal of the Responder\u0026rsquo;s displeasure over the violation of an implied expectation of fair treatment. Studies using the UG have consistently revealed that, indeed, a higher proportion of Responders reject disadvantageous splits (typically starting around the 30%Responder-70%Proposer range) than would be predicted by orthodox, rational choice theories (Nowak et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Morewedge et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). In this respect, the UG differs from the related Dictator Game (DG); in the DG, the Responder is not afforded the opportunity to reject the offer.\u003c/p\u003e\u003cp\u003eTo what extent does this pattern extend to human-AI interaction? In one of the first studies to investigate the UG in this context, Sanfey et al. (2003) reported from a small sample (19 participants) that Responders reacted with more emotion (likely anger) in response to a disadvantageous offer from a human than from a computer counterpart. Participants\u0026rsquo; anterior insula activation during the rejection of inequitable offers correlated with emotional expressions, suggesting a possible neural substrate for the relevant emotional processes. The authors proposed that participants reacted more negatively toward the human\u0026rsquo;s disadvantageous offer than the computer\u0026rsquo;s disadvantageous offer because people grant a higher degree of autonomy (Plaks et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), agency (Oudah et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), intentionality (Ayad \u0026amp; Plaks, 2025), and, in turn, moral responsibility (e.g. Strohminger \u0026amp; Jordan, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) to humans than to AI agents. (See also van 't Wout et al., \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2006\u003c/span\u003e for related findings.)\u003c/p\u003e\u003cp\u003e\u003cb\u003eEvolving Expectations of AI Agents\u003c/b\u003e\u003c/p\u003e\u003cp\u003eIn the intervening decades, however, enormous advances in computational power and neural architecture have led to the proliferation of commercially available AI products. This has brought human-AI interaction more squarely into the realm of everyday experience. As such, expectations about AI moral behavior are likely to have evolved. The idea of anthropomorphically blaming, for example, ChatGPT (not the programmers, not the OpenAI corporation) for a perceived transgression may seem less far-fetched to laypeople in 2025 than in the past.\u003c/p\u003e\u003cp\u003eFor this reason, we hypothesized a reversal of the \u003cem\u003eSanfey et al. effect in the Ultimatum Game\u003c/em\u003e: When provided with a disadvantageous offer, present-day people will respond more negatively toward AI agents than toward humans. We suspected this for four reasons. First, recent studies have demonstrated that people generally hold machines (relative to humans) to - and expect - a higher standard of reliability, a phenomenon dubbed the Perfect Automation Schema (Merritt et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; see also Oudah et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Shariff et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Thus, a violation of such expected reliability and fairness represents a greater betrayal (Bohnet \u0026amp; Zeckhauser, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2004\u003c/span\u003e). Second, many people may hold the assumption that AI agents should \u003cem\u003eserve\u003c/em\u003e humans, not undermine them (e.g., de Freitas et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). A disadvantageous offer may violate the AI agent\u0026rsquo;s presumed subordinate role. Third, interpersonal politeness norms that demand negative emotion suppression (Smith, Young, \u0026amp; Ford, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) may be reduced or absent when interacting with an AI counterpart (compared to a human counterpart). That is, people may feel freer to express their outrage toward an entity that is assumed to lack the capacity to experience feelings, such as insult or hurt (Oudah et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Fourth, participants may feel less reluctant to aggress toward, or even \u0026lsquo;punish\u0026rsquo;, an AI counterpart because they see it as a mechanism that possesses fewer (if any) moral rights (Strohminger \u0026amp; Jordan, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Treiman et al., \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cb\u003eAdvantageous Offers Are Also a Violation of Fairness Norms\u003c/b\u003e\u003c/p\u003e\u003cp\u003eConsiderably less work in this area has examined the psychology of receiving an unfairly advantageous offer, or \u003cem\u003eadvantageous inequity\u003c/em\u003e. Our method permitted us to measure participants\u0026rsquo; offer rejection rate, self-reported affect, and heart-rate variability in response to the experience of being over benefitted (e.g., 80%Responder-20%Proposer) by either a human or an AI counterpart. While a strict, rational self-interest approach might suggest that people should enjoy receiving unexpectedly high returns, there are grounds to suspect that, in practice, people will generally find advantageous inequity discomfiting. First, people may experience feelings of guilt upon receiving a payout that is perceived to be larger than deserved. This concept has a long history in psychological research, including classic work in Equity Theory in social relationships (Walster, 1978), as well as relatively recent conceptualizations (Bolton \u0026amp; Ockenfels, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Dunning et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Second, people may doubt their ability to maintain the positive outcomes into the future and may thus experience anxiety over the likely impending decrease (e.g., Plaks \u0026amp; Stecher, 2007). Third, separately from the valence of the payout (i.e., over- vs. under benefit), people may generally experience confusion upon receiving any sort of unexpected result. Finally, the literature on downregulation of positive emotion (e.g., Schall et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Zou, Plaks, \u0026amp; Peterson, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) indicates that, upon receiving positive news, people are typically sensitive to the emotional state of others in their environment and calibrate their own emotional expression so as not to deviate too much in a positive direction.\u003c/p\u003e\u003cp\u003eStudies on \u003cem\u003eAdvantageous Inequity Aversion\u003c/em\u003e (AIA) have demonstrated that medial frontal negativity (MFN), an index of neural activity associated with expectancy violation, responds negatively to advantageous inequities (Oliveira et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Wu and Zhou, \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). This aversion to unfair advantage aligns with societal norms that equate fairness with equal treatment, an observation supported by behavioral responses in the Ultimatum Game, in which advantageous offers are often rejected despite the potential gain (McAuliffe et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Furthermore, research with children indicates that this aversion is socially conditioned and develops with age. Whereas younger children tend to accept unfair advantages, by around age 8 they begin to reject such benefits, prioritizing fairness over personal gain (Blake \u0026amp; McAuliffe, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; McAuliffe et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Finally, there is evidence that people tend to view inequities (whether advantageous or disadvantageous) as violations of moral principles (Shaw \u0026amp; Choshen-Hillel, \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTo what extent might such concerns extend to AI agents? Given that most people assume that AI entities are (a) largely incapable of experiencing emotions (Oudah et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) and (b) less socially embedded than humans (Ayad \u0026amp; Plaks, 2025), these interpersonal emotional concerns should be largely moot when one is over benefitted by an AI agent. Thus, we hypothesized that participants would display lower advantageous inequity aversion (expressed by lower rejection rates, lower emotional disturbance, and higher heart-rate variability) when over benefitted by an AI counterpart than when over benefitted by a human counterpart.\u003c/p\u003e\u003cp\u003eOur design involved a 21-round game, rather than one shot. This allowed us to aggregate across participants\u0026rsquo; responses to multiple fair or unfair offers for arguably a more robust measure than can be provided by a single, one-shot game. In addition, the extended play period permitted us to measure physiological indices of heart rate variability. We suspected that such indices, which we describe next, would provide additional insight into the processes that contribute to accept/reject decisions.\u003c/p\u003e\u003cp\u003e\u003cb\u003eHeart Rate Variability\u003c/b\u003e\u003c/p\u003e\u003cp\u003eA healthy heart does not beat like a regular clock. Instead, its oscillations are complex and non-linear. Heart rate variability (HRV) refers to the fluctuation in the time interval between adjacent heartbeats (McCraty et al., 2015). It indexes neurocardiac function and is generated by heart-brain interactions and dynamic, non-linear autonomic nervous system (ANS) processes (Shaffer et al., 2017). The beat-to-beat fluctuations of a healthy heart are best described by mathematical chaos (Goldberger, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e1991\u003c/span\u003e). Such nonlinear variability is thought to provide the flexibility to rapidly cope with an uncertain and changing environment (Beckers et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). An optimal level of HRV is associated with health, self-regulatory capacity, and adaptability or resilience (McCraty et al., 2015).\u003c/p\u003e\u003cp\u003eHRV (measured via time, frequency, and non-linear metrics) has been identified as a physiological marker of stress-induced elevation in the sympathetic nervous system (SNS) and reduction in the parasympathetic nervous system (PNS) (Kim et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Time-domain indices of HRV quantify the amount of variability in measurements of the inter-beat interval (IBI). Frequency-domain measurements estimate the distribution of absolute or relative power into frequency bands. Non-linear measurements allow researchers to quantify the unpredictability of a time series (Stein \u0026amp; Reddy, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2005\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eSeveral studies have focused on how individual differences in heart rate variability are associated with decision-making in the UG. For example, S\u0026uuml;tterlin et al. (\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2011a\u003c/span\u003e) reported that participants with higher resting HRV, which indicates greater parasympathetic nervous system activity and inhibitory control, were more likely to reject unfair offers in the UG. Additionally, performance on a separate motor response inhibition task (Stop Signal Task, SST) also predicted rejection rates in the UG. Combining HRV and SST measures explained a significant portion of the variance in rejection rates, suggesting that self-regulatory capacity plays a crucial role in overcoming economic self-interest and promoting fairness-related behavior in the UG. In a related study, Dulleck et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) reported that lower HRV (indicative of higher self-regulatory effort and/or stress) correlated with both rejecting \u003cem\u003eand making\u003c/em\u003e unfair offers in the UG.\u003c/p\u003e\u003cp\u003eThe primary time-domain measure used to estimate vagally mediated changes reflected in HRV is the root mean square of successive differences between normal heartbeats (RMSSD) (Shaffer et al., 2017). Osumi and Ohira (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) found that heart rate deceleration was more pronounced when participants received offers that they subsequently rejected, suggesting that the perception of disadvantage triggers physiological responses that contribute to the decision to reject. Another research team reported that criminal judges were not only slower in their responses but also rejected a greater proportion of unfair offers, with their decision-making and rejection rates significantly correlated with higher HRV scores (Santamar\u0026iacute;a-Garc\u0026iacute;a, H. et al., 2021). These data suggest that HRV reflects self-regulatory effort in moral decision-making contexts, including the UG context. Taken together, these findings have introduced a new perspective to UG behavior by highlighting inhibitory control processes that may contribute to the decision to reject an offer.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec2\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003eOne hundred and fifty participants initially enrolled in the study. However, due to participant nonresponse or system malfunctions, the final sample size for analysis was 142 participants (Mean age\u0026thinsp;=\u0026thinsp;19.22, SD\u0026thinsp;=\u0026thinsp;2.68)\u003csup\u003e2\u003c/sup\u003e, with 63 randomly assigned to the human counterpart condition and 79 to the AI counterpart condition. The sample size was determined using a priori power analysis, informed by effect sizes and sample sizes reported in comparable prior studies (e.g., S\u0026uuml;tterlin et al., 2011; Santamar\u0026iacute;a-Garc\u0026iacute;a et al., 2021). The study was approved by the Research Ethics Board of the University of Toronto. All participants provided informed consent before the experimental procedures. Socio-demographic data (age, race/ethnicity, socioeconomic status) were collected via end questionnaires. Exclusion criteria included cardiovascular disease, arrhythmias, serious medical conditions, and the use of heart rate variability-affecting medications (e.g., beta-blockers, calcium channel blockers). Participants also reported their frequency of alcohol intake, tobacco intake, caffeine intake, minimum fifteen-minute exercise intervals, and minimum fifteen-minute meditation intervals on a weekly basis, due to their impact on HRV. Participants were required to abstain from smoking and caffeine on the day of the experiment and to avoid strenuous exercise and heavy meals prior to testing. The data supporting the findings of this study are available upon request from the lead author.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eProcedure\u003c/h2\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e After providing informed consent, proposers and responders briefly introduced themselves. In the AI condition, participants were greeted by \"SAM,\" embodied as a small, disk-shaped audio speaker. In reality, a research assistant fulfilled this role remotely, via the \u0026lsquo;Wizard of Oz\u0026rsquo; method; Unbeknownst to participants, SAM\u0026rsquo;s utterances were pre-recorded using Amazon Polly\u0026rsquo;s text-to-voice function and the sound files were played through the speaker by an assistant in an adjoining room. Thus, we wish to note that although we use the term \u0026lsquo;AI\u0026rsquo; to refer to SAM, it was not, in fact, an artificially intelligent agent. However, post-study queries revealed that all participants reported that they \u003cem\u003ebelieved\u003c/em\u003e SAM to be an AI agent. SAM greeted the participants with, \u0026ldquo;Hi, my name is SAM, it\u0026rsquo;s nice to meet you today, how have you been doing\u0026rdquo;? SAM was always presented as the proposer. In the Human condition, the proposer - a research assistant posing as another participant - made a similar verbal greeting. Next, participants completed a series of questionnaires, most of which test hypotheses that are beyond the scope of the present article and will thus not be discussed further and can be found in the supplementary material \u003csup\u003e3\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThen each participant was taken to a separate room and connected to the physiological recording system. Throughout the session, continuous heart rate monitoring was conducted using PPG sensors. (See details below.) After tests of the PPG system, participants turned their attention to the computer terminal to play a modified version of the Ultimatum Game (UG). To briefly recap the UG procedure, on each round, the proposer offers a split of points ranging from 0 to 100 (e.g., 70/30, such that the proposer keeps 70 and 30 goes to the responder). If the responder (the participant) agrees to the division, both parties receive the proposed amounts. However, a rejection by the responder results in neither player earning any points. All participants were assigned the role of the responder after completing prerequisite comprehension questions about the game rules. Research assistants, posing as fellow participants, posed as proposers in the human counterpart condition. The sequence of offers was quasi-randomized to ensure that each participant received an equal distribution of offers\u0026mdash;7 fair, 7 disadvantageous, and 7 advantageous.\u003c/p\u003e \u003cp\u003eOn each round, the proposer\u0026rsquo;s offer was presented to participants on the screen. Participants\u0026rsquo; task was to select \u0026lsquo;Accept\u0026rsquo; or \u0026lsquo;Reject\u0026rsquo;. The Ultimatum Game was conducted using the oTree platform (Chen, Schonger, \u0026amp; Wickens, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Following each of the 21 rounds, participants completed a brief, self-reported affect questionnaire - modeled on Osumi and Ohira (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). In keeping with the exact language used by Osumi and Ohira, participants rated their feelings of \u003cem\u003eanger\u003c/em\u003e, \u003cem\u003eaversion\u003c/em\u003e, \u003cem\u003ereassurance\u003c/em\u003e, and \u003cem\u003epleasure\u003c/em\u003e. In addition, we added a new exploratory item (\u003cem\u003edisgust\u003c/em\u003e) in light of the extensive literature on the link between disgust and moral condemnation (e.g., Inbar, Pizarro, \u0026amp; Bloom \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Robinson, Xu, \u0026amp; Plaks, \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eHeart Rate Variability Measurement\u003c/h3\u003e\n\u003cp\u003eTo measure heart rate variability, we used photoplethysmogram (PPG), which detects cardiovascular pulse waves via a light source and detector. This technique captures variations in blood volume or flow, with the intensity of backscattered light corresponding to changes in blood volume (Evans and Geddes \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e1988\u003c/span\u003e; Alnaeb et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). PPG offers a low-cost, non-invasive alternative to the traditional electrocardiogram (ECG). Studies have demonstrated PPG's effectiveness in monitoring healthy individuals, suggesting its viability as a replacement for ECG in some settings (Bolanos et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). Leveraging embedded devices like Raspberry Pi and Arduino increases the functionality of PPG. Consistent with Borthakur (2020), we used the Pulse Sensor Amped, an Arduino-based heart-rate sensor, to collect PPG data from participants, recording from the initial baseline to the final baseline. Previous data, including analyses in both time and frequency domains, indicate that PPG can be as reliable as ECG when a sampling frequency of at least 25 Hz is used (Choi and Shin \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). To optimize accuracy, participants were asked to minimize movement of the hand equipped with the sensor. We used the Python toolbox NeuroKit2 (Makowski, 2021) for data analysis. The raw PPG signal, sampled at 160.414 Hz, underwent preprocessing with a bandpass filter (0.5\u0026ndash;8 Hz, Butterworth 3rd order, following Elgendi et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). For more accurate RR interval estimation from PPG signals, we deployed specific functions from NeuroKit2 based on methodologies reported by Orphanidou et al. (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) and Frasch, M. G. (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eData Analysis\u003c/h2\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003eAccept/Reject Decisions\u003c/h2\u003e \u003cp\u003eBehavioral data were processed using R Studio Version 2024.4.2.764. We defined rejection rates as the \u003cem\u003eratio of rejected offers to total offers in each category\u003c/em\u003e (Disadvantageous, Advantageous, Fair, and Overall). Offer types were categorized as Disadvantageous if the participant was offered 0\u0026ndash;40%, Fair if offered 50%, and Advantageous if offered 60\u0026ndash;100%. Given non-normality of the distribution of rejection rates, a Kruskal-Walli\u0026rsquo;s test was conducted to compare rejection rates between different offer types. The Wilcoxon Signed-Rank Test was used to assess differences in rejection rates between the Human and AI conditions for advantageous and disadvantageous offers. Consistent with previous approaches in the literature (e.g., Osumi \u0026amp; Ohira (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), repeated measures ANOVAs were performed on the affect questionnaire data to evaluate the effect of offer type (Fair, Advantageous, Disadvantageous) on subjective feelings. Additionally, Spearman\u0026rsquo;s rank correlation was used to examine relationships between self-reported emotions, heart rate variability (HRV) metrics, and rejection rates and counts. To ensure consistency across affective measures, negative affects (Anger, Aversion, and Disgust) were reverse coded on a standardised scale, so that higher values indicated higher positive affect. Although \u0026lsquo;anger,\u0026rsquo; \u0026lsquo;aversion,\u0026rsquo; \u0026lsquo;reassurance,\u0026rsquo; and \u0026lsquo;pleasure\u0026rsquo; suggest four semantically distinct emotion concepts, analyses revealed that participants\u0026rsquo; responses were highly intercorrelated: Cronbach\u0026rsquo;s \u0026#120514; = 0.86. Given this high degree of intercorrelation, we averaged the scores for each participant to create a composite affect index, with higher values indicating higher affect positivity.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e\n\u003ch3\u003ePhysiological data analysis\u003c/h3\u003e\n\u003cp\u003eFrom the PPG recordings, we extracted the following time, frequency, and nonlinear measures using the neurokit2 toolbox:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eStandard Deviation of NN Intervals (SDNN), which indicates overall heart rate variability.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eLow Frequency (LF) and High Frequency (HF) components, which reflect a combination of sympathetic and parasympathetic activity, and parasympathetic activity, respectively.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eCardiac Vagal Index (CVI), which specifically measures parasympathetic function.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eMeanNN represents the mean of all normal-to-normal (NN) intervals in milliseconds, which is inversely related to heart rate.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eRoot Mean Square of the Successive Differences (RMSSD), which captures short-term variability linked to parasympathetic activity.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003epNN50, which is the percentage of NN intervals differing by more than 50 ms.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eShannon Entropy, which measures the complexity and unpredictability of heart rate patterns.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eSD1 and SD2, which reflect short-term and long-term autonomic regulation, respectively, and are considered to be sensitive to dynamic changes in stress and recovery.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eWe calculated these measures during the baseline period (5 min), then estimated it over three five-minute recording windows during the Ultimatum game play. The first and last 30 seconds of the data were excluded from the analysis to avoid potential artifacts or signal instability commonly observed during the start and end of PPG/ECG recordings, ensuring more reliable and accurate measurements. For short-term HRV measurements, 5-minute segments are generally considered appropriate, along with shorter durations (Cammet al. 1996).\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e1. Behavioral results: Accept/Reject Decisions\u003c/h2\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e1.1 Rejection Rates Varied by Offer Type\u003c/h2\u003e \u003cp\u003eFirst, to ascertain that the data replicated previous findings (e.g., Osumi and Ohira (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2009\u003c/span\u003e); Haselhuhn and Mellers, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; van't Wout et al., \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2006\u003c/span\u003e), we compared the rejection rates by offer type (Advantageous, Fair, Disadvantageous), collapsing across the human and AI counterpart conditions. A Kruskal-Walli’s test revealed a significant difference in rejection rates among the different offer types in the predicted direction (i.e. highest for Disadvantageous, lowest for Advantageous), χ² (2) = 257, p \u0026lt; .001.\u003c/p\u003e \u003cp\u003e \u003cem\u003e1.2 The Rejection Rate for Disadvantageous Offers Was Higher in the AI Condition Than in the Human Condition\u003c/em\u003e \u003c/p\u003e \u003cp\u003eNext, we compared the rejection rates for disadvantageous offers between the human and AI conditions. A Wilcoxon rank-sum test revealed that participants were more likely to reject disadvantageous offers from an AI counterpart than from a human counterpart, W = 1688, p = .0067 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Note that this is consistent with our hypothesis; whereas Sanfey et al. (2003) reported a higher rejection rate with the human counterpart, these data - collected over two decades later and with a much larger sample - indicate a reversal of the original effect.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e1.3 The Advantageous Offer Rejection Rate Was Higher in the Human Condition:\u003c/h2\u003e \u003cp\u003eNext, we compared the rejection rates for advantageous offers between the human and AI conditions. A Wilcoxon rank-sum test indicated that participants exhibited a significantly higher rejection rate for advantageous offers when playing with a human counterpart, W = 2703.5, p = .007 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). This is also consistent with our hypothesis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2. Impact of Affect\u003c/h2\u003e \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e \u003ch2\u003e1.4 Subjective Affect Varied by Human vs AI Condition\u003c/h2\u003e \u003cp\u003eAfter removing mindless or logically incoherent entries (for more information on exclusion criteria, see Supplementary Material \u003csup\u003e4\u003c/sup\u003e), we examined the effect of counterpart type and offer type on self-reported affect following fair offers, disadvantageous offers, and advantageous offers. The results are summarised in Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, 4 and 5. Recall that higher values indicate higher affect positivity.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e​​Disadvantageous Offers\u003c/h2\u003e \u003cp\u003eA one-way analysis of variance (ANOVA) on the self-reported affect index following disadvantageous offers revealed a significant effect of Counterpart Type (Human vs. AI), F (1, 128) = 6.79, p = 0.01, ηp² = 0.05. The direction of the effect indicates that participants reported feeling less positive affect / more negative affect following disadvantageous offers made by AI counterparts than by human counterparts.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eAdvantageous Offers\u003c/h2\u003e \u003cp\u003eIn contrast, an analogous ANOVA on self-reported affect following advantageous offers revealed no significant effect of Counterpart Type, F (1, 130) = 0.707, p = 0.402, ηp² = 0.005. This suggested no substantial difference in self-reported affect when participants received advantageous offers from human versus AI counterparts.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eFair Offers:\u003c/h2\u003e \u003cp\u003eAn analogous ANOVA for fair offers revealed no significant main effect of Counterpart Type (Human vs. AI) on self-reported affect, F (1, 130) = 1.981, p = 0.162, ηp² = 0.015. This indicates that participants’ emotional responses to fair offers did not differ significantly when the offers were made by a human counterpart or an AI counterpart.\u003c/p\u003e \u003cp\u003eThese results suggest that AI counterparts only elicit more negative affect than human counterparts following a disadvantageous offer. This pattern is consistent with the finding that participants were more likely to reject disadvantageous offers from the AI counterpart than from the human counterpart.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e1.5 Subjective Affect Was Generally Correlated with Rejection Rate, But Did Not Vary as a Function of Counterpart Type (AI versus Human).\u003c/em\u003e \u003c/p\u003e \u003cp\u003eTo assess the relationship between subjective affect and rejection rate, we conducted correlation analyses between affective responses and rejection rates for each offer type, as depicted in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eDisadvantageous Offers:\u003c/h2\u003e \u003cp\u003eFollowing disadvantageous offers, affect positivity was negatively correlated with rejection rate in both the AI condition (rs = − 0.321, p \u0026lt; 0.01) and the Human condition (rs = − 0.497, p \u0026lt; 0.001). To investigate whether the relation between affect and rejection rate was stronger in the Human condition than in the AI condition, we regressed rejection rate onto affect positivity, counterpart type and the interaction term. Results indicated that affect positivity significantly predicted rejection rate, B = -0.089, SE = 0.022, p \u0026lt; .001, with the negative sign indicating that more positive affect was associated with lower rejection rate. However, the interaction between counterpart type and affect was not significant, B = 0.030, SE = 0.030, p = .325, suggesting that the relationship between affect and rejection rate did not differ as a function of playing with an AI or human counterpart.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e\u003cem\u003eAdvantageous Offers\u003c/em\u003e:\u003c/h2\u003e \u003cp\u003eFor advantageous offers, affect positivity was negatively correlated with rejection in the AI condition (rs = − 0.291, p \u0026lt; 0.05) and in the Human condition (rs = − 0.465, p \u0026lt; 0.001). As above, we performed regression analysis to investigate whether participants experienced stronger emotional responses when rejecting advantageous offers in the Human condition compared to the AI condition. Results indicated a significant main effect of counterpart type, B = -0.071, SE = 0.028, p = .013, suggesting that participants rejected advantageous offers from human counterparts more frequently than from AI counterparts. Affect also significantly predicted rejection rate, B = -0.084, SE = 0.018, p \u0026lt; .001, with higher positive affect associated with lower rejection rates. The interaction between counterpart type and affect approached significance, B = 0.042, SE = 0.024, p = .078, indicating that, if anything, the effect of affect on rejection rate was stronger in the human condition, although this association should be interpreted with caution.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e\u003cem\u003eFair Offers\u003c/em\u003e:\u003c/h2\u003e \u003cp\u003eAffect positivity following fair offers was negatively correlated with rejection rate of fair offers in the AI condition, (rs = − 0.339, p \u0026lt; 0.01) and the Human condition (rs = − 0.477, p \u0026lt; 0.001). That is, the more negative (less positive) affect participants reported feeling after an offer, the more likely they were to reject that offer. To test for possible differences between the AI condition and the human condition, we conducted a regression analysis analogous to the ones above. It revealed that the main effect of counterpart type was not significant, B = -0.046, SE = 0.035, p = .194, indicating no significant difference in Fair offer rejection rates between AI and Human counterparts. However, affect significantly predicted rejection rate, B = -0.075, SE = 0.014, p \u0026lt; .001, with more positive affect associated with lower rejection rates. The interaction between counterpart type and affect was not significant, B = 0.027, SE = 0.024, p = .254, suggesting that the effect of affect on rejection rate did not differ between AI and Human counterpart.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCorrelations between affect and offer rejection rates in the human and AI conditions. Significance levels: *p \u0026lt; .05, **p \u0026lt; .01, ***p \u0026lt; .001.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCounterpart Type\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFair RR\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAdvantageous RR\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDisadvantageous RR\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAI\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.339**\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.291*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.321**\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eH\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.477***\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.465***\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.497***\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003eIn summary, analyses of participants’ self-reported affect revealed that participants reported feeling (a) generally more displeasure following disadvantageous offers than following the other types of offers and (b) more displeasure following disadvantageous offers from an AI counterpart than from a human counterpart. However, these differences did not differentially influence participants’ ultimate decisions to accept or reject offers from AI versus human counterparts. This may indicate that self-reported affective state is a comparatively crude measure, susceptible to both reporting biases and high variability. It may also indicate that additional processes, beyond affect play an important role in accept/reject decisions. We turn next to such potential processes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e2. Heart Rate Variability\u003c/h2\u003e \u003cp\u003e \u003cem\u003e2.1 Manipulation Check: Analysis of HRV Responses Across Initial Baseline, Ultimatum Game, and Final Baseline Periods\u003c/em\u003e \u003c/p\u003e \u003cp\u003eFirst, to examine the effect of simply playing the Ultimatum Game on Heart Rate Variability (HRV), we analyzed HRV across three time periods: the initial baseline (B1), the UG period, and the final baseline (B2). Significant changes in HRV were observed across these periods, reflecting physiological variations. Because no significant interaction or main effect of counterpart type (Human vs. AI) was detected, we combined the conditions (H/AI) and conducted a repeated measures one way ANOVA to focus on the overall effect of time. For analysis, for consistency we focused on the middle five minutes of the UG condition as baseline and UG was of different length. Data from 12 participants were excluded due to missing the final baseline or both the ultimatum game (UG) and the final baseline caused by sensor data collection errors.\u003c/p\u003e \u003cp\u003eA one-way ANOVA revealed a significant main effect of time for CVI, F (1.74, 190.09) = 17.76, p \u0026lt; .001, partial η² = .14. Pairwise comparisons demonstrated that CVI was significantly lower during the initial baseline compared to the second baseline, t (109) = − 5.08, p \u0026lt; .001, Cohen’s d = − 0.48. The initial baseline was not significantly different from the UG period, t (109) = − 2.35, p = .062, Cohen’s d = − 0.22. Additionally, CVI was significantly higher during the second baseline compared to the UG period, t (109) = 4.34, p \u0026lt; .001, Cohen’s d = 0.41. For RMSSD, a significant main effect of time was observed, F (1.66, 181.02) = 4.30, p = .015, partial η² = .038. Pairwise comparisons showed no significant difference between the initial and second baselines, t (109) = − 2.33, p = .066, Cohen’s d = − 0.22, nor between the initial baseline and the UG period, t (109) = − 0.71, p = .481, Cohen’s d = − 0.07. However, RMSSD was significantly higher in the second baseline compared to the UG period, t (109) = 2.53, p = .039, Cohen’s d = 0.24. For SD1, a significant main effect of time was observed, F (1.66, 181.03) = 4.31, p = .015, partial η² = .038. Pairwise comparisons indicated no significant difference between the initial baseline and the second baseline, t (109) = − 2.33, p = .065, Cohen’s d = − 0.22, or between the initial baseline and the UG period, t (109) = − 0.71, p = .481, Cohen’s d = − 0.07. The second baseline was significantly higher than the UG period, t (109) = 2.53, p = .039, Cohen’s d = 0.24. For SD2, a significant main effect of time was observed, F (1.83, 199.98) = 29.17, p \u0026lt; .001, partial η² = .211. Pairwise comparisons revealed that SD2 was significantly lower during the initial baseline compared to the second baseline, t (109) = − 6.57, p \u0026lt; .001, Cohen’s d = − 0.63. The initial baseline was significantly lower than the UG period, t (109) = − 2.55, p = .037, Cohen’s d = − 0.24. Additionally, SD2 was significantly higher during the second baseline compared to the UG period, t (109) = 5.34, p \u0026lt; .001, Cohen’s d = 0.51. For SDNN, a significant main effect of time was observed, F (1.78, 194.42) = 16.49, p \u0026lt; .001, partial η² = .131. Pairwise comparisons indicated that SDNN was significantly lower during the initial baseline compared to the second baseline, t (109) = − 5.44, p \u0026lt; .001, Cohen’s d = − 0.52. The initial baseline was not significantly different from the UG period, t (109) = − 2.01, p = .142, Cohen’s d = − 0.19. SDNN was significantly higher during the second baseline compared to the UG period, t(109) = 4.87, p \u0026lt; .001, Cohen’s d = 0.46. For Shannon Entropy, a significant main effect of time was observed, F (1.78, 194.42) = 21.25, p \u0026lt; .001, partial η² = .163. Pairwise comparisons revealed that ShanEn was significantly lower during the initial baseline compared to the second baseline, t (109) = − 4.85, p \u0026lt; .001, Cohen’s d = − 0.46. The difference between the initial baseline and the UG period was not significant, t (109) = − 1.97, p = .154, Cohen’s d = − 0.19. However, ShanEn was significantly higher in the second baseline compared to the UG period, t (109) = 4.24, p \u0026lt; .001, Cohen’s d = 0.40. Lower HRV during the initial baseline aligns with anticipatory stress before engaging in a task (Pulopulos et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The UG period likely imposed cognitive or emotional challenges, reflected in reduced HRV measures such as SDNN, CVI, RMSSD, SD1, SD2. Similarly, Entropy reductions during stress periods (initial baseline and UG) likely reflect lower complexity in heart rate variability, a marker of reduced autonomic adaptability under stress (Pincus \u0026amp; Goldberger, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e1994\u003c/span\u003e). Increased variability during the second baseline indicates the individual's recovery and adaptive capacity following the stressor, as suggested by research on autonomic recovery dynamics (Kim et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThese analyses provide evidence that HRV measures do capture heightened cognitive or emotional demands during the UG period. These data also serve as a manipulation check. The initial baseline's lower HRV likely reflects anticipatory stress, while the final baseline's higher HRV indicates recovery and relaxation following task completion.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Association between Heart Rate Variability and Ultimatum Game Decisions\u003c/h2\u003e \u003cp\u003eNext, we examined the relationship between various HRV measures and fair, advantageous, disadvantageous, and overall offer rejection rate in the Human and AI conditions. Spearman correlations are presented in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, as they are generally robust to outliers (Croux \u0026amp; Öllerer, 2015). The outliers were not removed, but rather winsorized following ​​the recommendations of Pulopulos, Vanderhasselt, and De Raedt (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The results reveal distinct patterns across various HRV metrics. In line with Sütterlin et al. (\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2011a\u003c/span\u003e), who reported Pearson correlations, we also tested Pearson correlations. Detailed results, largely align with the Spearman correlation findings.\u003c/p\u003e \u003cp\u003eAt the aggregate level, a correlational analysis suggests differences in HRV metrics between AI and Human conditions. For overall rejection rate, correlations between HRV measures and rejection rate were consistently stronger in the AI condition: SDNN (0.334 for AI vs. 0.062 for H), SD2 (0.401 for AI vs. 0.098 for H), and ShanEn (0.312 for AI vs. 0.035 for H). These differences indicate that participants in the AI condition generally displayed a stronger association between physiological metrics and rejection of offers.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eDisadvantageous Offers\u003c/h2\u003e \u003cp\u003eFor disadvantageous offers, a similar trend emerged, with stronger correlations in the AI condition for SD2 (0.201 for AI vs. -0.08 for H) and ShanEn (0.153 for AI vs. -0.114 for H). The negative correlations in the Human condition (e.g., MeanNN: -0.167 for H vs. 0.181 for AI) suggest differing physiological responses to unfairness when interacting with AI versus humans. To examine more directly whether the type of counterpart (human vs. AI) influenced the relationship between heart rate variability (HRV) predictors and rejection of disadvantageous offers, we conducted multiple regression analyses. We found no correlation between body mass index (BMI) and HRV, so we excluded it as a confounding variable in the subsequent regression analysis. We regressed the disadvantageous offer rejection rate onto each specific HRV measure, condition (Human vs. AI), and the interaction term. Significant interaction effects were observed for SDNN (β = -0.0028, \u003cem\u003ep\u003c/em\u003e = .021), MeanNN (β = -0.0011, \u003cem\u003ep\u003c/em\u003e = .002), SD2 (β = -0.0023, \u003cem\u003ep\u003c/em\u003e = .040), and ShanEn (β = -0.1652, \u003cem\u003ep\u003c/em\u003e = .003), such that participants in the AI condition exhibited a stronger positive association between these HRV metrics and rejection rate than did participants in the human condition.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eAdvantageous Offers\u003c/h2\u003e \u003cp\u003eFor advantageous offers, the differences between AI and Human conditions were less pronounced but still notable. Measures such as SD2 (0.161 for AI vs. 0.102 for H) and MeanNN (0.185 for AI vs. 0.212 for H) exhibited comparable correlations in both conditions, reflecting similar physiological responses to favorable outcomes. To examine more directly whether the type of counterpart (Human vs. AI) influenced the relationship between heart rate variability (HRV) measures and rejection of advantageous offers, we conducted analogous regression analyses. Rejection rate was regressed onto each HRV measure, Counterpart Type (Human vs. AI), and the interaction term. No significant interaction effects were observed for SDNN (β = -0.00005, \u003cem\u003ep\u003c/em\u003e = .940), MeanNN (β = -0.00001, \u003cem\u003ep\u003c/em\u003e = .947), SD2 (β = -0.00009, \u003cem\u003ep\u003c/em\u003e = .892), or ShanEn (β = -0.0011, \u003cem\u003ep\u003c/em\u003e = .974), indicating that the associations between these HRV measures and rejection rates did not differ between the Human and AI conditions. A significant main effect of MeanNN was observed, β = 0.0004, \u003cem\u003ep\u003c/em\u003e = .006, suggesting a positive association between MeanNN and rejection rate regardless of whether the counterpart was a human or an AI agent. Additionally, counterpart type significantly predicted rejection rates, β = -0.052, \u003cem\u003ep\u003c/em\u003e = .005, with participants in the Human condition rejecting advantageous offers more frequently than participants in the AI condition. Neither Affect (β = -0.002, \u003cem\u003ep\u003c/em\u003e = .813) nor its interaction with SDNN (β = -0.0002, \u003cem\u003ep\u003c/em\u003e = .327) were significant predictors of rejection rate.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eFair Offers\u003c/h2\u003e \u003cp\u003eFor fair offers, the Spearman correlation analysis revealed stronger associations in the AI condition compared to the Human condition for most HRV measures. Notable differences include SDNN (0.377 for AI vs. 0.068 for H), SD2 (0.405 for AI vs. 0.099 for H), and ShanEn (0.344 for AI vs. 0.077 for H). To examine more directly whether the type of counterpart (Human vs. AI) influenced the relationship between heart rate variability (HRV) measures and rejection of fair offers, we conducted analogous multiple regression analyses. Rejection rate was regressed onto each HRV measure, Counterpart Types (Human vs. AI), and their interaction term. Significant interaction effects were observed for SDNN (β = -0.0023, \u003cem\u003ep\u003c/em\u003e = .016) and SD2 (β = -0.0021, \u003cem\u003ep\u003c/em\u003e = .014), indicating that these HRV measures were more strongly associated with rejection rate in the AI condition compared to the human condition.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\"Comparison of Spearman correlations between HRV measures (SDNN, SD1, SD2, RMSSD, MeanNN, ShanEn, and CVI) and Ultimatum Game decisions (rejection rate for Fair offers, Disadvantageous offers, and Overall Rejection Rate) in the Human and AI conditions. Significant differences highlight stronger associations in the AI condition across multiple HRV metrics.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMeasure\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFair (H)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFair (AI)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDisadv (H)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDisadv (AI)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAdv (H)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAdv (AI)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eOverall RR\u003c/p\u003e \u003cp\u003e(H / AI)\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSDNN\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.068\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.377***\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.106\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.160*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.122\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.138\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.062 / 0.334***\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSD1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.005\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.338***\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.074\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.145\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.069\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.069\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.056 / 0.264***\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSD2\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.099\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.405***\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.080\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.201**\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.102\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.161*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.098 / 0.401***\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRMSSD\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.005\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.338***\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.074\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.145\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.069\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.069\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.056 / 0.264***\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMeanNN\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.074\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.077\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.167*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.181*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.212*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.185*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e-0.031/ 0.257***\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eShanEn\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.077\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.344***\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.114\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.153*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.095\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.096\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.035 / 0.312***\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCVI\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.022\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.340***\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.088\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.157*\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.120\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.107\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.054 / 0.294***\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003eIn summary, these analyses reveal stronger HRV associations with rejection rate in the AI condition than in the Human condition, suggesting that individuals experience heightened physiological activation when interacting with an AI agent (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e7\u003c/span\u003e). The inhibitory component of rejecting a disadvantageous offer is thought to involve overriding one’s default preference to accept any money that is offered (Sutterlin et al., 2011). Thus, these analyses suggest that participants’ higher rejection rate of disadvantageous offers from the AI counterpart is driven, in part, by higher parasympathetic system activity associated with inhibitory control. Was this association moderated by participants’ affect? If so, it would suggest that emotional tone contributes to shaping participants’ deployment of self-regulatory processes. If not, it would suggest that affective processes and self-regulatory processes represent largely independent contributions to accept/reject decisions. We turn next to this question.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn regression analyses of rejection rates of Advantageous, Disadvantageous, and Fair offers with Affect, HRV (SDNN), and the interaction term as predictors, distinct patterns emerged between participants who faced human versus AI counterparts. For Advantageous offers, participants in the human condition showed no significant main effects for SDNN (β = 0.00078, p = 0.220) or Affect (β = -0.00476, p = 0.541), but a significant interaction between SDNN and Affect (β = -0.00060, p = 0.034), suggesting that as parasympathetic activity (SDNN) increases, the relationship between Affect and rejection rate weakens or reverses. In contrast, participants in the AI condition displayed no significant main effects of SDNN (β = 0.00060, p = 0.156) or Affect (β = 0.00102, p = 0.843), and no interaction effects (β = 0.00011, p = 0.623).\u003c/p\u003e \u003cp\u003eFor Disadvantageous offers, participants in the human condition exhibited a significant main effect of Affect (β = -0.0266, p = 0.023), indicating that greater negative affect predicted higher rejection rates, while SDNN (β = -0.00057, p = 0.549) and the interaction term (β = -0.00005, p = 0.898) were non-significant. In the AI condition, in contrast, SDNN showed a significant positive main effect (β = 0.00174, p = 0.033), suggesting that higher parasympathetic activity was associated with increased inhibitory control when rejecting disadvantageous offers, whereas Affect (β = -0.0115, p = 0.252) and the interaction term (β = 0.00018, p = 0.675) were not significant.\u003c/p\u003e \u003cp\u003eFor Fair offers, participants in the human condition displayed a significant main effect of Affect (β = -0.0286, p = 0.008), indicating that greater negative affect was associated with higher rejection rates, while SDNN (β = 0.0012, p = 0.159) and the interaction (β = -0.00050, p = 0.184) were not significant. In the AI condition, SDNN emerged as a significant predictor of rejection rate (β = 0.00279, p \u0026lt; 0.001), while Affect (β = -0.0064, p = 0.310) and the interaction term (β = -0.00025, p = 0.365) were non-significant.\u003c/p\u003e \u003cp\u003eIn summary, these analyses suggest distinct mechanisms driving rejection decisions when interacting with human versus AI counterparts. Specifically, the decision to reject an offer appears to be influenced by two sources: (a) affective experience and (b) inhibition of the default tendency to accept any amount of money. With a human counterpart, the decision to reject an offer is primarily driven by affective state, such that higher positive emotion reduces the likelihood of rejection. This reflects a relatively straightforward process, as participants can readily interpret the experience of receiving a disadvantageous offer within a typical, human-to-human, normative framework. In contrast, when interacting with an AI counterpart, rejection decisions are less influenced by emotional states, but more influenced by the deployment of inhibitory processes—reflected in higher parasympathetic activity (e.g. SDNN)—which override the default tendency to accept any offer. This additional inhibitory control may stem from participants' unfamiliarity and uncertainty with how to navigate social conventions with a quasi-human entity, making the act of rejecting an AI offer comparatively more effortful. Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e8\u003c/span\u003e illustrates this proposed hydraulic relation between affect and self-regulation by enlarging the affect component for human counterparts, but enlarging the self-regulation component for AI counterparts.\u003c/p\u003e "},{"header":"General Discussion","content":"\u003cp\u003eThese data yielded several novel findings regarding individuals’ behavioral, physiological, and affective responses to disadvantageous, advantageous, and fair offers from AI agents. First, participants were more likely to reject disadvantageous offers from an AI counterpart than from a human counterpart. This pattern directly contrasts with findings reported by Sanfey et al. (2003) (and was obtained with a much larger sample). It is attributable, we suggest, to at least four reasons. First, with the passage of time, AI entities have assumed a larger role in millions of people’s everyday lives. This has likely led to evolving expectations of AI social behavior since 2003. Second, research on the Perfect Automation Schema (Merrit et al., 2015) has demonstrated that people often expect higher reliability and impartiality from autonomous agents. For example, Shariff, Bonnefon, and Rahwan (\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) reported that participants held autonomous vehicles to a higher standard of safety than they required from human drivers. Third, people may believe that AI agents should be subservient to humans (De Freitas et al, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2023\u003c/span\u003e); thus, an AI agent that attempts to domineer a human with a lowball offer would violate its presumed subservient position. These findings align with Treiman et al. (\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), who demonstrated that participants even exhibit a willingness to incur personal costs to instill fairness in AI. Fourth, although LLMs are increasingly convincing as conversational agents, the social norms that inhibit negative emotional expressions in human-to-human interactions may nonetheless be weaker in human-to-AI interactions (Oudah et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Thus, people may feel freer to express displeasure toward - or even ‘punish’ - an entity that will presumably not take offence.\u003c/p\u003e\u003ch2\u003eInfluence of HRV versus Affect\u003c/h2\u003e\u003cp\u003eAlthough participants generally reported feeling worse after receiving disadvantageous offers, self-reported positive/negative affect was equally associated with the rejection of disadvantageous offers from human and AI counterparts. In contrast, the relation between heart rate variability (HRV) metrics such as MeanNN, SDNN, and RMSSD and rejection rate displayed a clear sensitivity to counterpart type (AI vs. Human), with stronger associations between parasympathetic engagement and rejection of offers from AI counterparts than from human counterparts. These HRV metrics are commonly thought to index individuals’ level of effortful focus. Thus, the experience of playing the Ultimatum Game with an AI agent, and receiving a lowball offer from an AI agent, appear to intensify self-control processes.\u003c/p\u003e\u003cp\u003eTaken together, we propose the following explanatory model (depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e8\u003c/span\u003e): In typical, human-human interactions, societal fairness norms act as a check against the default tendency to accept any money that is offered. Such culturally ingrained fairness norms contribute to the rejection of both disadvantageous offers (“You are ripping me off!”) and advantageous offers (“I am ripping you off!”). In human-AI interactions, however, such social norms have yet to be clearly established and are thus weaker. In such cases, to restrain the default inclination to accept a disadvantageous offer, additional input is needed. This input appears to come from effortful, self-regulatory processes, captured by measures of HRV. In other words, in the absence of traditional emotional cues, physiological regulation reflected in HRV may serve as an important complementary mechanism that contributes to the decision to accept or reject the offer. This model aligns with prior research that has emphasized the role of HRV in regulating fairness-related behavior (Sütterlin et al., 2011).\u003c/p\u003e\u003ch2\u003eAdvantageous Inequity Aversion (AIA):\u003c/h2\u003e\u003cp\u003eAdvantageous offers, despite being beneficial, are perceived as violations of fairness norms, leading to discomfort linked to guilt or embarrassment. Participants in this study demonstrated higher rejection rates of advantageous offers from human counterparts than from AI counterparts. This finding resonates with work by Shaw and Choshen-Hillel (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) and McAuliffe et al. (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), who provided evidence for the role of societal norms and social conditioning in shaping fairness perceptions.\u003c/p\u003e\u003cp\u003eOne possible contributor to this phenomenon is emotional synchrony. This phenomenon of emotional synchrony was captured by Adam Smith in his \u003cem\u003eTheory of Moral Sentiments\u003c/em\u003e, in which he describes how individuals moderate their emotional expressions (including positive emotions) to align with societal expectations:\u003c/p\u003e\u003cp\u003e \u003cem\u003e“To see the emotions of their hearts, in every respect, beat in time to his own constitute his sole consolation. But he can only hope to obtain this by lowering his passion to that pitch in which the spectators are capable of going along with him. He must flatten, if I may be allowed to say so, the sharpness of its natural tone…”\u003c/em\u003e [Smith, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e1759\u003c/span\u003e\u003cem\u003e/2009, p. 28].\u003c/em\u003e\u003c/p\u003e\u003cp\u003eThat is, people often “flatten” their positive emotional tone when interacting with other humans, to avoid eliciting resentment or discomfort. This emotional calibration, however, becomes less relevant in interactions with AI agents, who are perceived, as the default, to be generally lower in - or devoid of - emotional capacity (Ayad \u0026amp; Plaks, 2025; Yam et al., \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Although the current procedure did not include any measures to directly test this positive emotion suppression mechanism, we consider it to be a plausible partial explanation for the diminished rejection rates of advantageous offers and the lower emotional disturbance when dealing with AI agents. We encourage future researchers to test this possibility more directly.\u003c/p\u003e\u003cp\u003eOur findings extend prior research by demonstrating the critical role of inhibitory control and emotional regulation, measured by HRV, in navigating social norms of fairness with AI agents. Previous studies (e.g., Sanfey et al., 2003; Knoch et al., 2006) highlighted the importance of executive functioning in regulating fairness-related behavior, often mediated by the dorsolateral prefrontal cortex. The present study suggests that such self-regulatory processes captured by HRV compensate for the absence of external social cues and well-established social norms in AI interactions. In short, whereas emotional responses play the dominant role in human-human UG exchanges, self-regulation appears to assume a larger role when the counterpart is an AI agent. We suggest that in future studies, HRV measures might emerge as effective indices for capturing decision-making in human-AI interactions, particularly in contexts involving fairness.\u003c/p\u003e\u003cp\u003eThe findings have significant implications for the design and implementation of AI systems in social and economic contexts. Understanding how people perceive fairness in AI interactions can guide the development of systems that align with human moral and ethical standards. For example, embedding human-like social cues in AI agents could mitigate the reliance on physiological regulation, fostering trust and cooperation (Plaks, Bustos-Rodriguez, \u0026amp; Ayad, 2022). Moreover, as AI becomes more integrated into daily life, shifting cultural norms surrounding human-AI interaction must be continually assessed. Finally, clear cultural differences in comfort with robotic and AI agents have been documented (Castelo \u0026amp; Sarvary, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, Lim et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Future researchers will need to be mindful of participants’ cultural context when assessing their perceptions of AI fairness.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eACKNOWLEDGMENTS:\u003c/h2\u003e \u003cp\u003e The authors acknowledge the valuable insights provided by members of the University of Toronto Motivation and Social Cognition laboratory. This work is partially funded by grants from the Natural Sciences and Engineering Research Council of Canada (fund # 508977) and the Schwartz-Reisman Institute for Technology and Society (SRI- VPRI).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAlnaeb, M. E., Alobaid, N., Seifalian, A. M., Mikhailidis, D. P., and Hamilton, G. (2007). Optical techniques in the assessment of peripheral arterial disease. Current vascular pharmacology, 5(1), 53\u0026ndash;59.\u003c/li\u003e\n \u003cli\u003eAyad, R. \u0026amp; Plaks, J.E. (2025). Attribution of intent and moral responsibility to AI agents.\u003cbr\u003e\u0026nbsp;Computers in Human Behavior: Artificial Humans, 3, 100107.\u003c/li\u003e\n \u003cli\u003eAxelrod, R. (2006). \u003cem\u003eThe Evolution of Cooperation\u003c/em\u003e (Revised ed.). Basic Books.\u003c/li\u003e\n \u003cli\u003eBeckers, F., Verheyden, B., \u0026amp; Aubert, A. E. (2006). Aging and nonlinear heart rate control in a healthy population. American Journal of Physiology-Heart and Circulatory Physiology, 290(6), H2560\u0026ndash;70. doi:10.1152/ajpheart.00903.2005\u003c/li\u003e\n \u003cli\u003eBishara, A. J., \u0026amp; Hittner, J. B. (2012). Testing the significance of a correlation with nonnormal data: comparison of Pearson, Spearman, transformation, and resampling approaches. Psychological methods, 17(3), 399.\u003c/li\u003e\n \u003cli\u003eBlake, P. R., \u0026amp; McAuliffe, K. (2011). \u0026ldquo;I had so much it didn\u0026rsquo;t seem fair\u0026rdquo;: Eight-year-olds reject two forms of inequity. Cognition, 120(2), 215-224.\u003c/li\u003e\n \u003cli\u003eBohnet, I., \u0026amp; Zeckhauser, R. (2004). Trust, risk and betrayal. Journal of Economic Behavior \u0026amp; Organization, 55(4), 467\u0026ndash;484. https://doi.org/10.1016/j.jebo.2003.11.004\u003c/li\u003e\n \u003cli\u003eBolanos, M., Nazeran, H., \u0026amp; Haltiwanger, E. (2006). Comparison of heart rate variability signal features derived from electrocardiography and photoplethysmography in healthy individuals. 2006 International Conference of the IEEE Engineering in Medicine and Biology Society, IEEE, 4289\u0026ndash;4294.\u003c/li\u003e\n \u003cli\u003eBolton, G. E., \u0026amp; Ockenfels, A. (2000). ERC: A theory of equity, reciprocity, and competition. \u003cem\u003eAmerican Economic Review\u003c/em\u003e, 90(1), 166\u0026ndash;193.\u003c/li\u003e\n \u003cli\u003eCastelo, N., \u0026amp; Sarvary, M. (2022). Cross-cultural differences in comfort with humanlike robots: Evidence from Japan and the United States. International Journal of Social Robotics.\u003c/li\u003e\n \u003cli\u003eChen, D. L., Schonger, M., \u0026amp; Wickens, C. (2016). \u003cem\u003eoTree\u0026mdash;An open-source platform for laboratory, online, and field experiments.\u003c/em\u003e Journal of Behavioral and Experimental Finance, 9(C), 88\u0026ndash;97.\u003c/li\u003e\n \u003cli\u003eChoi, A., \u0026amp; Shin, H. (2017). Photoplethysmography sampling frequency: Pilot assessment of how low can we go to analyze pulse rate variability with reliability? Physiological measurement, 38(3), 586.\u003c/li\u003e\n \u003cli\u003eChristophe, C., \u0026amp; \u0026Ouml;llerer, V. (2015). Robust and sparse estimation of the inverse covariance matrix using rank correlation measures. Social Science Research Network, 35-55. doi:10.1007/978-81-322-3643-6_3\u003c/li\u003e\n \u003cli\u003eCroux, C., \u0026amp; \u0026Ouml;llerer, V. (2016). Robust and sparse estimation of the inverse covariance matrix using rank correlation measures (pp. 35-55). Springer India.\u003c/li\u003e\n \u003cli\u003eDe Freitas, J., Agarwal, S., Schmitt, B., \u0026amp; Haslam, N. (2023). Psychological factors underlying attitudes toward AI tools.\u0026nbsp;Nature Human Behaviour,\u0026nbsp;7(11), 1845-1854.\u003c/li\u003e\n \u003cli\u003eDunning, D., Anderson, J. E., Schl\u0026ouml;sser, T., Ehlebracht, D., \u0026amp; Fetchenhauer, D. (2014). Trust at zero acquaintance: more a matter of respect than expectation of reward.\u0026nbsp;Journal of Personality and Social Psychology,\u0026nbsp;107(1), 122.\u003c/li\u003e\n \u003cli\u003eDulleck, U., Schaffner, M., \u0026amp; Torgler, B. (2011). Heartbeat and economic decisions: Observing mental stress among proposers and responders in the ultimatum bargaining game. PLOS ONE.\u003c/li\u003e\n \u003cli\u003eElgendi, M., Norton, I., Brearley, M., Abbott, D., \u0026amp; Schuurmans, D. (2013). Systolic peak detection in acceleration photoplethysmograms measured from emergency responders in tropical conditions.\u0026nbsp;PloS one,\u0026nbsp;8(10), e76585.\u003c/li\u003e\n \u003cli\u003eEvans, A. M., \u0026amp; Revelle, W. (2008). Survey and behavioral measurements of interpersonal trust. \u003cem\u003eJournal of Research in Personality\u003c/em\u003e, 42(6), 1585\u0026ndash;1593.\u003c/li\u003e\n \u003cli\u003eEvans, M., \u0026amp; Geddes, L. (1988). An assessment of blood vessel vasoactivity using photoplethysmography. Medical instrumentation, 22(1), 29\u0026ndash;32.\u003c/li\u003e\n \u003cli\u003eFrasch, M. G. (2022). Comprehensive HRV estimation pipeline in Python using Neurokit2: Application to sleep physiology. MethodsX, 9, 101782.\u003c/li\u003e\n \u003cli\u003eGoldberger, A. L. (1991). Is the normal heartbeat chaotic or homeostatic? News Physiol Sci, 6, 87\u0026ndash;91.\u003c/li\u003e\n \u003cli\u003eG\u0026uuml;th, W., Schmittberger, R., \u0026amp; Schwarze, B. (1982). An experimental analysis of ultimatum bargaining. \u003cem\u003eJournal of Economic Behavior and Organization\u003c/em\u003e, 3(4), 367\u0026ndash;388.\u003c/li\u003e\n \u003cli\u003eHarsanyi, J. C. (1961). On the rationality postulates underlying the theory of cooperative games. \u003cem\u003eJournal of Conflict Resolution\u003c/em\u003e, 5(2), 179\u0026ndash;196.\u003c/li\u003e\n \u003cli\u003eHaselhuhn, M.P., Mellers, B.A., 2005. Emotions and cooperation in economic games. Brain Res. Cogn. Brain Res. 23, 24\u0026ndash;33.\u003c/li\u003e\n \u003cli\u003eInbar, Y., Pizarro, D. A., \u0026amp; Bloom, P. (2009). Conservatives are more easily disgusted than liberals. Cognition and Emotion, 23(4), 714\u0026ndash;725.\u003c/li\u003e\n \u003cli\u003eJean-Fran\u0026ccedil;ois Bonnefon, Iyad Rahwan, and Azim Shariff, \u0026ldquo;The Moral Psychology of Artificial Intelligence\u0026rdquo;, Annual Review of Psychology, vol. 75, January 2024, pp. 653\u0026ndash;675.\u003c/li\u003e\n \u003cli\u003eJohnson-George, C., \u0026amp; Swap, W. C. (1982). \u003cem\u003eMeasurement of Specific Interpersonal Trust: Construction and Validation of a Scale to Assess Trust in a Specific Other\u003c/em\u003e. Journal of Personality and Social Psychology, 43(6), 1306\u0026ndash;1317.\u003c/li\u003e\n \u003cli\u003eKarpus, J., Kr\u0026uuml;ger, A., Verba, J. T., Bahrami, B., \u0026amp; Deroy, O. (2021). Algorithm exploitation: Humans are keen to exploit benevolent AI.\u0026nbsp;Iscience,\u0026nbsp;24(6).\u003c/li\u003e\n \u003cli\u003eKim, H. G., Cheon, E. J., Bai, D. S., Lee, Y. H., \u0026amp; Koo, B. H. (2018). Stress and heart rate variability: a meta-analysis and review of the literature. Psychiatry investigation, 15(3), 235.\u003c/li\u003e\n \u003cli\u003eKosfeld, M., Heinrichs, M., Zak, P. J., Fischbacher, U., \u0026amp; Fehr, E. (2005). Oxytocin increases trust in humans. Nature, 435(7042), 673\u0026ndash;676.\u003c/li\u003e\n \u003cli\u003eLim, A., Louie, W.-Y. G., \u0026amp; Su, N. M. (2021). Social robots on a global stage: Establishing a role for culture during human\u0026ndash;robot interaction. International Journal of Social Robotics, 13(4), 1023\u0026ndash;1037.\u003c/li\u003e\n \u003cli\u003eMakovi K, Sargsyan A, Li W, Bonnefon JF, Rahwan T. 2023. Trust within human-machine collectives depends on the perceived consensus about cooperative norms. Nature Communications\u003c/li\u003e\n \u003cli\u003eMakowski, D., Pham, T., Lau, Z. J., Brammer, J. C., Lespinasse, F., Pham, H., ... \u0026amp; Chen, S. A. (2021). NeuroKit2: A Python toolbox for neurophysiological signal processing. Behavior research methods, 1-8.\u003c/li\u003e\n \u003cli\u003eMalle, B. F., \u0026amp; Ullman, D. (2021). A multidimensional conception and measure of human-robot trust. In Trust in human-robot interaction (pp. 3-25). Academic Press.\u003c/li\u003e\n \u003cli\u003eMcAuliffe, K., Blake, P. R., Kim, G., Wrangham, R. W., \u0026amp; Warneken, F. (2013). Social influences on inequity aversion in children. PLOS ONE, 8(12), e80966.\u003c/li\u003e\n \u003cli\u003eMcAuliffe, K., Blake, P. R., \u0026amp; Warneken, F. (2014). Children reject inequity out of spite. Biology letters, 10(12), 20140743.\u003c/li\u003e\n \u003cli\u003eMcCraty, R., \u0026amp; Shaffer, F. (2015). Heart rate variability: new perspectives on physiological mechanisms, assessment of self-regulatory capacity, and health risk. Global Advances in Health and Medicine, 4, 46\u0026ndash;61. doi:10.7453/gahmj.2014.073\u003c/li\u003e\n \u003cli\u003eMorewedge, C. K., Krishnamurti, T., Ariely, D. (2014). \u0026quot;Focused on fairness: Alcohol intoxication increases the costly rejection of inequitable rewards\u0026quot;. Journal of Experimental Social Psychology. 50: 15\u0026ndash;20. doi:10.1016/j.jesp.2013.08.006\u003c/li\u003e\n \u003cli\u003eMerritt, S. M., Unnerstall, J. L., Lee, D., \u0026amp; Huber, K. (2015). Measuring individual differences in the perfect automation schema.\u0026nbsp;Human factors,\u0026nbsp;57(5), 740-753.\u003c/li\u003e\n \u003cli\u003eNielsen YA, Pfattheicher S, Keijsers M. 2022a. Prosocial behavior toward machines. Current Opinion in Psychology 43:260\u0026ndash;265\u003c/li\u003e\n \u003cli\u003eNowak, M. A. , Page, K.M., \u0026amp; Sigmund, K. (2000). \u0026quot;Fairness Versus Reason in the Ultimatum Game\u0026quot;. Science. 289 (5485): 1773\u0026ndash;1775.\u003c/li\u003e\n \u003cli\u003eNowak, M. A., \u0026amp; Sigmund, K. (2005). Evolution of indirect reciprocity. \u003cem\u003eNature\u003c/em\u003e, 437(7063), 1291\u0026ndash;1298.\u003c/li\u003e\n \u003cli\u003eOliveira, F. T., McDonald, J. J., \u0026amp; Goodman, D. (2007). Performance monitoring in the anterior cingulate is not all error related: expectancy deviation and the representation of action-outcome associations. Journal of Cognitive Neuroscience, 19(12), 1994-2004.\u003c/li\u003e\n \u003cli\u003eOrphanidou, C., Bonnici, T., Charlton, P., Clifton, D., Vallance, D., \u0026amp; Tarassenko, L. (2015). \u003cem\u003eSignal-Quality Indices for the Electrocardiogram and Photoplethysmogram: Derivation and Applications to Wireless Monitoring\u003c/em\u003e. IEEE Journal of Biomedical and Health Informatics, 19(3), 832\u0026ndash;838.\u003c/li\u003e\n \u003cli\u003eOsumi, T., \u0026amp; Ohira, H. (2009). Cardiac responses predict decisions: An investigation of the relation between orienting response and decisions in the ultimatum game. Biological Psychology, 82(2), 174-181.\u003c/li\u003e\n \u003cli\u003eOudah, M., Makovi, K., \u0026amp; Gray, K. (2024). Perception of experience influences altruism and perception of agency influences trust in human\u0026ndash;machine interactions. \u003cem\u003eScientific Reports\u003c/em\u003e, 14, 63360.\u003c/li\u003e\n \u003cli\u003ePettit, P. (2018). \u003cem\u003eThe Birth of Ethics: Reconstructing the Role and Nature of Morality\u003c/em\u003e. Oxford University Press.\u003c/li\u003e\n \u003cli\u003ePincus, S. M., \u0026amp; Goldberger, A. L. (1994). Physiological time-series analysis: What does regularity quantify? American Journal of Physiology-Regulatory, Integrative and Comparative Physiology, 266(4), R1395-R1407.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePlaks, J. E., Rodriguez, L. B., \u0026amp; Ayad, R. (2022). Identifying psychological features of robots that encourage and discourage trust.\u0026nbsp;Computers in Human Behavior,\u0026nbsp;134, 107301.\u003c/li\u003e\n \u003cli\u003ePulopulos, M. M., Vanderhasselt, M. A., \u0026amp; De Raedt, R. (2018). Association between changes in heart rate variability during the anticipation of a stressful situation and the stress-induced cortisol response. Psychoneuroendocrinology, 94, 63-71.\u003c/li\u003e\n \u003cli\u003eRobinson, J.S., Xu, X. \u0026amp; Plaks, J.E. (2019). \u0026nbsp;Disgust and deontology: \u0026nbsp;Trait sensitivity to pathogens promotes a preference for clarity, hierarchy, and rule-based moral judgment. Social Psychological and Personality Science, 10\u003cem\u003e,\u0026nbsp;\u003c/em\u003e3-14\u003cem\u003e.\u003c/em\u003e\u003c/li\u003e\n \u003cli\u003eRosas, A., Berm\u0026uacute;dez, J. P., Cotrina, J. M., Aguilar-Pardo, D., Caicedo, J. C., \u0026amp; Aponte-Canencio, D. M. (2020). Perceiving utilitarian gradients: Heart rate variability and self-regulatory effort in the moral dilemma task. Journal of Moral Education, 49(1), 45-61.\u003c/li\u003e\n \u003cli\u003eRotter, J. B. (1967). A new scale for the measurement of interpersonal trust. \u003cem\u003eJournal of Personality\u003c/em\u003e, 35(4), 651\u0026ndash;665.\u003c/li\u003e\n \u003cli\u003eSantamar\u0026iacute;a-Garc\u0026iacute;a, H., Cotrina, J. M., Torres, N. F., Buitrago, C., Aponte-Canencio, D. M., Caicedo, J. C., Billeke, P., Gantiva, C., \u0026amp; Baez, S. (2020). Explicit and implicit markers of fairness preeminence in criminal judges. Scientific Reports, 10, 11234.\u003c/li\u003e\n \u003cli\u003eSchall, M., Martiny, S. E., Goetz, T., \u0026amp; Hall, N. C. (2016). Smiling on the inside: The social benefits of suppressing positive emotions in outperformance situations. Personality and Social Psychology Bulletin, 42, 559\u0026ndash;571.\u003c/li\u003e\n \u003cli\u003eShaffer, F., \u0026amp; Ginsberg, J. P. (2017). An overview of heart rate variability metrics and norms. Frontiers in Public Health, 5, 258.\u003c/li\u003e\n \u003cli\u003eShariff, A., Bonnefon, J.J. \u0026amp; Rahwan, I. (2021). How safe is safe enough? Psychological mechanisms underlying extreme safety demands for self-driving cars, Transportation Research Part C: Emerging Technologies, 126, 103069.\u003c/li\u003e\n \u003cli\u003eShaw, A., \u0026amp; Choshen-Hillel, S. (2017). It\u0026rsquo;s not fair: Folk intuitions about disadvantageous and advantageous inequity aversion. Judgment and Decision Making, 12(3), 208-223.\u003c/li\u003e\n \u003cli\u003eSheldon, O., Plaks, J.E., Sridharan, V., \u0026amp; Shoda, Y. (2018). Strategic actors in situ impressions of systematically- versus unsystematically-variable counterparts. Social Cognition, 36, 324-344.\u003c/li\u003e\n \u003cli\u003eSloane, S., Baillargeon, R., \u0026amp; Premack, D. (2012). Do infants have a sense of fairness? Psychological Science, 23(2), 196\u0026ndash;204.\u003c/li\u003e\n \u003cli\u003eSmith, A.M., Young, G. \u0026amp; Ford, B.Q. (2023). The interpersonal correlates of believing emotions are controllable. Motivation and Emotion, 47, 323\u0026ndash;332.\u003c/li\u003e\n \u003cli\u003eSmith, A. (1759). \u003cem\u003eThe Theory of Moral Sentiments\u003c/em\u003e.\u003c/li\u003e\n \u003cli\u003eStein, P. K., \u0026amp; Reddy, A. (2005). Non-linear heart rate variability and risk stratification in cardiovascular disease. Indian Pacing Electrophysiology Journal, 5, 210\u0026ndash;20.\u003c/li\u003e\n \u003cli\u003eStrohminger, N., \u0026amp; Jordan, M. R. (2022). Corporate insecthood. Cognition, 224, 105068.\u003c/li\u003e\n \u003cli\u003eS\u0026uuml;tterlin, S., Herbert, C., Schmitt, M., K\u0026uuml;bler, A., \u0026amp; V\u0026ouml;gele, C. (2011a). Overcoming selfishness: reciprocity, inhibition, and cardiac-autonomic control in the ultimatum game. Frontiers in Psychology, 2, 11221.\u003c/li\u003e\n \u003cli\u003eS\u0026uuml;tterlin, S., Herbert, C., Schmitt, M., K\u0026uuml;bler, A., \u0026amp; V\u0026ouml;gele, C. (2011b). Frames, decisions, and cardiac\u0026ndash;autonomic control. Social Neuroscience, 6(2), 169-177.\u003c/li\u003e\n \u003cli\u003eTreiman, L. S., Ho, C. J., \u0026amp; Kool, W. (2023, November). Humans forgo reward to instill fairness into AI. In Proceedings of the AAAI Conference on Human Computation and Crowdsourcing (Vol. 11, No. 1, pp. 152-162).\u003c/li\u003e\n \u003cli\u003evan \u0026apos;t Wout, M., Kahn, R. S., Sanfey, A. G., \u0026amp; Aleman, A. (2006). Affective state and decision-making in the Ultimatum Game. Experimental Brain Research, 169(4), 564-8. doi: 10.1007/s00221-006-0346-5.\u003c/li\u003e\n \u003cli\u003eWalster, E., Walster, G. W., \u0026amp; Berscheid, E. (1978). \u003cem\u003eEquity: Theory and research\u003c/em\u003e. Allyn \u0026amp; Bacon\u003c/li\u003e\n \u003cli\u003eWaytz, A., Cacioppo, J., \u0026amp; Epley, N. (2010). \u003cem\u003eWho Sees Human? The Stability and Importance of Individual Differences in Anthropomorphism\u003c/em\u003e. Perspectives on Psychological Science, 5(3), 219\u0026ndash;232.\u003c/li\u003e\n \u003cli\u003eWu, Y., \u0026amp; Zhou, X. (2009). The P300 and reward valence, magnitude, and expectancy in outcome evaluation. Brain Research, 1286, 114-122.\u003c/li\u003e\n \u003cli\u003eYam, K. C., Bigman, Y. E., Tang, P. M., Ilies, R., De Cremer, D., Soh, H., \u0026amp; Gray, K. (2020). Robots at work: People prefer\u0026mdash;and forgive\u0026mdash;service robots with perceived feelings. Journal of Applied Psychology, 106(10), 1557\u0026ndash;1572.\u003c/li\u003e\n \u003cli\u003eZou, C., Plaks, J.E., \u0026amp; Peterson, J.B. (2019). Don\u0026rsquo;t get too excited: Assessing individual differences in the downregulation of positive emotions. Journal of Personality Assessment, 101, 73-83.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e Such games raise a reasonable question: What would an AI agent do with any money it earns? The data thus far, however, indicate that participants are generally untroubled by this question (e.g., Makovi et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Plaks et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), likely assuming that although the money may be useless to the agent, it nonetheless has been programmed to attempt to maximize its earnings (Bonnefon et al., 2024)..\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e The age data were available for 136 participants.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e One measure assessed trait-level dehumanization tendencies (Yam et al., 2021). A second measure assessed individual differences in anthropomorphism tendencies (Waytz et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). A third measure was a single item assessing the perceived importance of fairness in decision-making. A fourth measure assessed trait-level interpersonal trust Johnson-George \u0026amp; Swap, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e1982\u003c/span\u003e).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e The results did not meaningfully vary whether the excluded values were included or excluded.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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":"","lastPublishedDoi":"10.21203/rs.3.rs-6051145/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6051145/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eHuman moral interactions often assume that resources should be allocated equitably, i.e., one should not take more than one\u0026rsquo;s fair share. To what extent do people apply this assumption to social AI entities? Using a 21-round Ultimatum Game, we investigated participants\u0026rsquo; behavioral, physiological, and affective responses to fair, disadvantageous, and advantageous offers from an AI (vs. human) counterpart. We report three principal findings: (a) Participants were more likely to reject disadvantageous offers from an AI counterpart than from a human counterpart, but were more likely to reject advantageous offers from a human counterpart than from an AI counterpart; (b) Participants reported more negative affect following disadvantageous offers from an AI counterpart than from a human counterpart; (c) Participants exhibited a stronger association between heart rate variability and rejection rate for disadvantageous offers from an AI counterpart than from a human counterpart. Based on these findings, we propose a model emphasizing an important, previously under-examined role of self-regulatory processes in humans\u0026rsquo; responses toward AI moral behavior.\u003c/p\u003e","manuscriptTitle":"Inequity Aversion Toward AI Counterparts","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-26 09:00:55","doi":"10.21203/rs.3.rs-6051145/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":"ed4b22e4-9292-4876-baa5-ea7ea291edd4","owner":[],"postedDate":"March 26th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":45194010,"name":"Social science/Psychology"},{"id":45194011,"name":"Biological sciences/Psychology"}],"tags":[],"updatedAt":"2025-03-26T09:00:57+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-26 09:00:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6051145","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6051145","identity":"rs-6051145","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}