Embedding Generative AI as a digital capability into a year-long MSc skills program

Embedding Generative AI as a digital capability into a year-long MSc skills program | 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 Research Article Embedding Generative AI as a digital capability into a year-long MSc skills program David P. Smith, Dami Sokoya, Skye Moore, Chinenya Okonkwo, Charlotte Boyd, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5204546/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 The arrival of Generative Artificial Intelligence (GenAI) into higher education has brought about significant transformations in assessment practices and pedagogical approaches. Large Language Models (LLMs) powered by GenAI present unprecedented opportunities for personalised learning journeys. However, the emergence of GenAI in higher education raises concerns regarding academic integrity and the development of essential cognitive and creative skills among students. Critics worry about the potential decline in academic standards and the perpetuation of biases inherent in the training sets used for LLMs. Addressing these concerns requires clear frameworks and continual evaluation and updating of assessment practices to leverage GenAI's capabilities while preserving academic integrity. Here, we evaluated the integration of GenAI into a year-long MSc program to enhance student understanding and confidence in using GenAI. Approaching GenAI as a digital competency, its use was integrated into core skills modules across two semesters, focusing on ethical considerations, prompt engineering, and tool usage. The assessment tasks were redesigned to incorporate GenAI, which takes a process-based assessment approach. Students' perceptions were evaluated alongside skills audits, and they reported increased confidence in using GenAI. Thematic analysis of one-to-one interviews revealed a cyclical relationship between students' usage of GenAI, experience, ethical considerations, and learning adaptation. Generative Artificial Intelligence Higher Education Process base Assessment Assessment AI Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Generative Artificial Intelligence (GenAI) has quickly become a transformative factor in higher education, influencing assessment practices and pedagogical approaches to learning and teaching. Large Language Models (LLMs) offer unprecedented capabilities for personalised learning journeys (Bearman & Luckin, 2020 ; Hooda et al., 2022 ). GenAI can deliver substantial efficiency gains and enhanced educational outcomes through personalised learning pathways when considered an enabler. However, the disruptive nature of the technology and its emergence in higher education has raised concerns about academic integrity, the authenticity of student work, and the potential reduction in student development of essential cognitive and creative skills (Richardson & Clesham, 2021 ; Tenakwah et al., 2023 ). Large Language Models (LLMs) are trained on large collections of text data to capture as much human language as possible (Rudolph et al., 2023 ). When asked questions, LLMs predict which words to respond with based on the training set. The GenAI recalls prior interactions during conversations to inform its next prediction, allowing for conversational interactions. Supporters of GenAI argue that it can enhance efficiency, productivity, and information access; however, critics are concerned that GenAI could lead to declining academic standards (Sallam, 2023 ). Due to the training sets used in creating LLMs, there is an inherent bias stemming from prejudices in the source material that reflect historical or social inequities. It is crucial to ensure that GenAI enhances human skills and understanding rather than replacing them (Bearman & Luckin, 2020 ; Bobula, 2024 ). This approach involves using clear frameworks and constantly evaluating and updating assessment practices to leverage GenAI's capabilities while maintaining the educational experience's integrity. GenAI has shown its rapidly evolving capabilities in producing realistic human-like text, convincing images, videos, and other media. Many evaluation methods rely on the capability to generate new content (the evaluation) based on a prompt (the evaluation brief) using existing knowledge (the training dataset). The emergence of AI-generated content presents clear challenges for preserving academic honesty in this context. Richardson and Clesham ( 2021 ) point out the difficulties in identifying AI-generated submissions, which can undermine the credibility of student work. Concerns about academic honesty in the context of artificial intelligence (AI) are supported by data showing challenges distinguishing between student-generated and AI-generated work (Elkhatat et al., 2023 ). There are also significant concerns that AI detectors unfairly put neurodiverse and non-native English language speakers at a disadvantage (Liang et al., 2023 ). The use of AI in education gives rise to various ethical concerns. It is crucial that AI systems are developed and implemented in a way that prevents the reinforcement of existing biases and promotes equal educational opportunities. Addressing these risks entails the establishment of clear guidelines for the use of AI in assessments (Perkins, 2023 ). The effective integration of AI tools into learning environments requires strategic planning and a commitment to pedagogical principles. GenAI tools have the potential to significantly enhance educational outcomes by providing personalised feedback, facilitating language learning, and supporting various research methodologies. Additionally, they can increase learner engagement and motivation while maintaining a strong ethical framework that focuses on privacy, bias, and accuracy. Understanding the impact of GenAI on education necessitates a comprehensive theoretical approach (Noroozi et al., 2024 ). Constructivism proposes that learning is an active process wherein learners build their understanding of the world through experiences and interactions with the environment (Fosnot, 2013 ). The integration of GenAI into this framework echoes social constructivism, where learning is viewed as an active process in which knowledge is constructed through social interactions (Zhou & Schofield, 2024 ). GenAI tools can facilitate collaborative learning and provide personalised feedback to enhance student engagement. Much of the power of GenAI lies in the creation of personalised learning environments through tools and systems that tailor educational experiences to individual student needs. These tools can dynamically adjust content delivery and assessment methods based on continuous data analysis, enhancing educational outcomes by aligning with individual learning trajectories. Competency-based learning assesses students’ ability to apply concepts in real-world scenarios rather than through traditional memory-based tasks (Huxley-Binns et al., 2023 ). Behaviourist principles, particularly reinforcement and feedback, are central to competency-based education. In this framework, learning is seen as a change in behaviour resulting from acquiring knowledge or skills, and students are assessed based on observable outcomes or competencies. This approach assesses students' ability to apply concepts in real-world scenarios. GenAI can facilitate competency-based learning by providing detailed feedback on performance and identifying areas for improvement, aiding learners in mastering specific competencies. In 2023, Moorhouse et al. ( 2023 ) reviewed guidelines from the top 50 universities in the Times Higher Education World University Rankings 2023. The review focused on the content and advice provided for instructors about using GenAI in assessments. The review encouraged redesigning assessments to incorporate GenAI tools effectively. This includes designing tasks with GenAI, emphasising critical thinking and creativity, and using in-class assessments to prevent potential misuse. Emphasising the learning process over final outputs and supporting staged assessments that allow for feedback and development were seen as good practices. AI's capabilities require innovative approaches to assessment design with a shift to creating assessments that demand critical thinking and a deeper understanding of the subject area (Bobula, 2024 ; Chan & Colloton, 2024 ). Assessment methods include open-ended questions (Meir et al., 2024 ), project-based assessments (Petrovska et al., 2024 ), and real-world problem-solving tasks that encourage original thought and application of knowledge. There is a need for a significant shift in thinking about assessment design and the iterative creation of the work to be submitted. Project-based learning addresses this need, where students must engage in a process over time, such as research projects, case studies, or design tasks, which requires original thinking and continuous instructor feedback (Petrovska et al., 2024 ). Developing assessments that explicitly require students to use AI tools for certain parts of the task, followed by a critical evaluation of the AI’s output, can help students develop critical thinking skills about AI and its capabilities (Sok & Heng, 2024 ). By using a product-oriented rather than a process-driven assessment model, it is more likely that even if students use GenAI to complete assessments, they will have to engage with, refine, and update the outputs of tools and evidence ethical usage of the AI package selected (Smith & Francis, 2024 ). Surveys of educators and students reveal diverse attitudes towards AI in education (Chan & Colloton, 2024 ; Lacey & Smith, 2023 ; Palmer et al., 2023 ). The analysis indicates a general recognition of AI's benefits in supplementing learning, assessment efficiency and feedback. However, there is also significant apprehension regarding the potential for academic dishonesty and its impact on learning processes. Many educators see the potential of AI in teaching but have concerns about ethical issues and academic integrity (Lodge et al., 2023 ). The use of generative AI in education challenges traditional practices and requires new approaches to foster creativity alongside the risk and the risk of undermining the authenticity of student work (Ali et al., 2024 ). Educators are also facing complexities and constraints in adapting these new digital tools into their practices and feel the need for AI literacy and support in this new digital competency. Student responses to GenAI in assessments vary. Chan & Colloton ( 2024 ) found that some students appreciate the personalised feedback and learning support that GenAI offers, while others worry about how it might affect their creativity and originality. These concerns highlight the importance of clear institutional policies and guidelines, as well as the necessity of educating students about ethical AI usage (Smith & Francis, 2024 ). Although students generally trust AI for tasks like grammar correction, they still prefer human educators for assessment feedback (Palmer et al., 2023 ; Smith & Francis, 2024 ). When comparing attitudes across different assessment situations, it is clear that tailored approaches to AI integration are needed to address specific worries and contexts. Recommendations include establishing comprehensive guidelines for AI use, ensuring that AI complements rather than replaces traditional teaching methods, and emphasising responsible and ethical implementation. By understanding and addressing these diverse perspectives, educational institutions can better navigate the challenges and opportunities brought by AI in assessment practices. While GenAI offers tremendous potential to enhance personalised learning experiences and educational outcomes, its integration into higher education must carefully consider academic integrity, ethical concerns, and ensuring that it complements, rather than replaces, essential human cognitive and creative skills. In this study, experience mapping, a concept developed by Colin Beard as part of his broader framework on experiential learning (Beard, 2022 ), was used to embed GenAI into a series of three core research skills modules. The process involves designing, analysing, and refining learning experiences to maximise their impact on learners, delivering and building core skills and competencies through structured delivery. Incorporating AI into module assessments in this study involved a shift towards emphasising the creation process rather than the product (Smith & Francis, 2024 ), especially with written articles (Rudolph et al., 2023 ). Assessing the process over the product has several advantages as it integrates tool usage as part of the overall output and is independent of the tool used, giving the assessment longevity (Romova & Andrew, 2011 ). Additionally, there are benefits to addressing academic integrity by structuring assessment tasks that contribute towards the final article (Sotiriadou et al., 2020 ) emphasises the learning journey and the steps taken rather than just evaluating the final product (Jeltova et al., 2007 ; Rodriguez et al., 2020 ; Rudolph et al., 2023 ; Schreiber et al., 2016 ). Here, we evaluate a year-long study in which GenAI was embedded into the skills development program as part of an MSc program. We aimed to determine if student understanding and confidence in the use of GenAI could be delivered and enhanced in a large cohort teaching environment. To achieve this, GenAI was integrated into core modules delivered each semester over a two-semester research skills program. Students reported an increase in confidence around GenAI. Thematic analysis of open-text response questions and interview transcripts describes a cyclical relationship between themes. As students use GenAI and gain more experience, they refine their usage, adapt their learning, and continually reassess the ethical implications. This, in turn, influences how they continue to use GenAI and the support they seek from it. Methods Curriculum delivery The experience mapping process utilised here involved designing, analysing, and refining learning experiences to maximise their impact on learners and was used to create a visual representation of the learning journey (Fig. 1 ). This map helps educators and students see the sequence of experiences, identifying critical touchpoints and was used to plan how different elements of GenAI skills development could be incorporated. During the first semester, points were identified as the need to understand GenAI's workings, the ethical and appropriate use of the tools, data integrity considerations and the institutional policies around academic conduct. These moments were crucial for understanding how AI produces its responses and reflected on the use of GenAI during their studies. During semester two, prompt design and example GenAI tools were covered that addressed specific needs, such as aiding in understanding, scaffold reflection, feeding forward on assessment tasks, or acting as a learning guide. During semester one, an introduction to GenAI and how it operates alongside the ethical and appropriate use of GenAI was delivered. During semester two, prompt engineering, examples and different tools were embedded into the taught delivery. Tool usage supported by video content and prompt design were appropriate to the topic and were delivered using the just-in-time principle, with example prompts supporting the activities of the week. These were presented to the students as the “GenAI Tip of the week,” delivered during seminars and supported by prompt libraries and video support via the virtual learning environment. Prompt Break down [topic you’d like to understand] into smaller, easier-to-understand parts. Use analogies and real-life examples to simplify and make the concept more relatable. Prompt How would you verify the information in this conversation? *Example prompt used to help understand more complex topics. GenAI prompts were also presented to facilitate the reflection process by analysing learner inputs and providing automated yet personalised feedback. Prompt As an MSc student enrolled on a Bioscience or Chemistry course in the UK, you require an action plan and tips to complete your next assignment. You have received feedback on your lab bookkeeping assessment which includes some example areas that need improvement. Please review the feedback and provide suggestions for enhancing the laboratory report assessment. *Example prompt used to gain feedforward advice following the initial assessment. Real-time Adaptation of Learning Paths was also presented, enabling real-time adaptation based on ongoing learner performance. GenAI prompts tools were presented that modify the learning path instantaneously, offering more challenges or support as needed. Prompt You are an [MSc Pharmaceutical Bioscience Student]. You are about to prepare a two-page CV to include with your application for a [PhD position]. Set out structure and ideas for content for an impactful CV.” Prompt In this conversation, you will take on the role of an interviewer. You are looking to hire an intern for one year. The pharmaceutical company is looking for a bench scientist. Ask questions that would be suitable for this role. You will ask the question, and I will then give you my answer. You will then give feedback. GenAI tools and prompts were also presented to help learners who struggled with a particular concept particularly understanding research articles or choosing the correct statistical method for a given situation. Prompts were provided to generating additional examples, exercises, or explanations to reinforce understanding. Prompt Your role in this conversation is to act as a guide helping a researcher to choose which statistical test to use when analysing their data. Your conversation will be based on working through a statistical decision tree. You will ask the researcher questions to guide them to the most appropriate statistical test. In your responses, give an explanation of the terms used below the main text. Use examples to help them understand. Your first question will be about the number of groups used in the study and ask for some background information. Are you ready? Assessment design incorporated GenAI prompts, ensuring the rationale for tool utilisation and critiquing source validity (Brew et al., 2023 ). Assessments were fully integrated into the taught context to help students develop these competencies and demonstrate learning outcomes. Seminars, and tutorials were conducted to foster the development of these competencies, including the ethical and proper use of GenAI, such as logging search strategies and keywords and critically appraising content validity (Crowther et al., 2010 ). GenAI was integrated into the assessment design by documenting prompts and justifying how the generated text contributed to the final article. Furthermore, templated guides were developed to structure this process, providing students with direction in documenting their progress towards the final product. A complete theoretical underpinning and assessment design strategy alongside assessment rubrics can be found in (Smith & Francis, 2024 ) A questionnaire was undertaken at the start of semester 1 to inform curriculum content based on students’ current understanding and perceptions of AI. Students’ self-reporting skills in a skills audit at the start of semester 1 and the end of semester 2 determined the impact of the teaching and learning within the curriculum. To allow for a more detailed understanding of the student experience, interviews with students were undertaken in semester 3. Participants The student participants were MSc cohorts on a range of Biosciences and Chemistry master’s degree programmes comprising Analytical Chemistry and Pharmaceutical Analysis. Biomolecular Science, Biotechnology and Pharmacology, Microbiology Biology and Cancer Biology. The cohort is drawn from various nationalities, primarily Nigeria, India, Pakistan and the United Kingdom. In all degree programmes, students participate in core skills modules that run for three semesters, with weekly tutorials and seminars. Seminar sessions are co-taught with the full cohort and are split into tutorial groups of 25–30 students. Laboratory sessions are a mandatory component of the degree programmes linked directly to assessment. From a total cohort of 180 students, 156 opted into the study (87% opt-in rate). Ethics Ethics for this study were acquired through the College of Health, Wellbeing, and Life Sciences ethics committee following the Sheffield Hallam University Research Ethics Policy (ER61054725). Ethical approval was given as no identifiable, confidential, or controversial information would be collected. No gender, age, educational experience or other demographic factors were requested or considered in the analysis. Participation in the study was optional. The following statement was displayed before in-class polling took place. We would like to evaluate your responses so that we can further improve our skills development program over the coming years specifically around AI and digital skills. Responses to the in-class poll are fully anonymised; no personally identifiable information will be collected. If you do not wish to take part please note your responses for your own records Before data collection for the skills audit, students were read and given a copy of the following statement, which served as a means of consent. “We would like to evaluate your responses so that we can further improve our skills development program over the coming years specifically around AI and digital skills. To do this we will collect your responses in semester 1 and semester 2 and ask the question how have your skills developed. We will use your SHU e-mail to track your responses and the details will only be available to the study organised by Professor David Smith. Taking part is entirely voluntary. Not taking part will not affect your studies or assessments. Input your student number if you consent to your anonymised responses being evaluated.” Students were recruited for the interviews via e-mails from the virtual learning environment and posters displayed in the teaching laboratories. As an incentive, a ten-pound gift card was given. Participants again gave consent by completing an interview consent form after reading the study information sheet (Supplementary Information). Evaluation Instruments Questionnaires The student questionnaires (Supplementary Information) were collected, and an independent researcher indexed and transposed the details and comments into Excel. Students' names in the study were replaced with an identifier to prevent linking the responses to an individual and bias by the investigator. Responses to Likert scale questions were converted to numbers from 1 to 5 for statistical analysis, with 1 representing the most negative response to the question (e.g. strongly disagree) and 5 the most positive (e.g. strongly agree). The ordinal nature of the questionnaire data meant that parametric statistics were inappropriate for analysis, so nonparametric tests were used throughout. Observations were independent, with no individuals belonging to more than one study group. For Mann–Whitney, U-tests between groups were used as appropriate. Statistical significance is indicated as P < 0.05, P < 0.01. Skills Audit Skills audits (supplementary information) were conducted early in semester one and at the end of semester two. Data was collected via an online questionnaire during a large group-taught session. Students were asked to evaluate the importance and confidence of a range of research skills, including GenAI, on a Linkert scale of 1 = not important/confident, 2 = Limited importance/confident, 3 = somewhat important/confident, 4 = quite important/confident, and 5 = extremely important/confident. A copy of the responses was e-mailed automatically to the students for their own records. Interviews This study employed a qualitative research design, utilising semi-structured interviews to explore participants’ experiences and perceptions of GenAI (Supplementary Information). Fifteen participants were recruited through email invitations. These participants represented a diverse cross-section of the MSc cohort. Each participant participated in a one-to-one, structured interview designed to elicit detailed responses related to their experiences with GenAI. Informed consent was obtained from all participants prior to the interview. Students were asked to confirm that they had read the participation sheet and consented to the interview. Automatic transcription built within teams was used to generate transcripts. To prepare the transcripts for analysis, an initial data cleaning process was undertaken. This involved the removal of redundant words, filler words (e.g., “um,” “like”), and other non-essential verbal tics, following best practices in qualitative data management (Gibbs, 2007). The cleaned transcripts kept the integrity of the participants’ responses while ensuring clarity for later thematic analysis. The data were analysed using thematic analysis, following the six-phase process outlined by Braun and Clarke (2019). Initially, the researchers familiarised themselves with the data by reading and re-reading the transcripts, followed by systematic coding of key phrases and patterns relevant to the research questions. These codes were grouped into broader themes, which were then reviewed and refined to ensure they accurately captured the participants’ perspectives. The themes were clearly defined and named, and the final analysis was written up with supporting quotations from the participants. Throughout the process, reflexivity was maintained, and steps were taken to ensure the rigour and trustworthiness of the analysis by following guidelines from Lincoln and Guba (1985). Results Evaluation of initial attitudes to GenAI use. To establish the student cohort's base-level understanding and attitudes to GenAI, students were polled during the initial GenAI seminar using a student response (clicker) system. The question followed taught content around how LLM’s and GenAI models operate. Students were asked, “How often do you use GenAI tools?” with 44% stating they use GenAI tools Always or Often (Fig. 2 ). An open-text response was then used to capture which tools or models the students were using with ChatGPT was identified as the primary GenAI tool used by the student cohort. A set of Likert scale questions was employed to gauge students' initial perceptions and attitudes towards using GenAI in their studies. Students were asked to indicate whether they believed the following scenarios were acceptable uses of GenAI tools. The survey assessed participants' views on AI's usefulness across various academic tasks, such as writing assessments, understanding content, editing written work, and locating research articles (Fig. 3 ). The questionnaire reveals varying confidence levels in AI's capabilities across these areas. Comprehension most respondents had a favourable view of AI's role in aiding content comprehension. 89% of participants agreed or strongly agreed that AI was helpful, with only 4% strongly disagreeing and 7% disagreeing. This suggests a strong confidence in AI's ability to assist in understanding complex topics. Research articles AI was overwhelmingly viewed as beneficial for finding and understanding research articles. A substantial 86% of respondents agreed or strongly agreed with this capability, with only a small minority − 6% strongly disagreeing and 8% disagreeing - expressing doubts. This finding highlights the perceived value of AI tools in streamlining the research process and improving access to scholarly materials. Overall, the data suggest that while there is strong confidence in AI's ability to assist with content comprehension and research, students understood the ethical concerns around using GenAI in generating and editing written assessments. Editing written work Opinions on AI's ability to help with editing written content were more divided. While 43% of respondents agreed or strongly agreed that AI could assist in editing, 28% strongly disagreed, and 29% disagreed. This spread indicates that, although some users find AI useful for editing, a significant proportion remains unconvinced of its effectiveness. Write assessments When asked if it was acceptable that AI could effectively write their assessments, a significant portion of respondents replied negatively, with 44% strongly disagreeing and 34% disagreeing. Only 22% of respondents agreed or strongly agreed that AI could perform assessment tasks. This highlighted that the majority of students understood concerns around academic integrity and highlighted considerable doubt about AI's ability to generate accurate and complete assessments independently. These data informed the content and tone of taught sessions within the module. Assessment of perceived skills development A skills audit was conducted at the beginning and end of the module to assess how students perceived their skills around AI had progressed. Initially, the audit provides a baseline understanding of each student’s perceived existing knowledge and competencies. At the end of the modules, a second audit measures the self-reported improvements or changes in the skill set, thereby highlighting growth areas and identifying remaining gaps (Fig. 4 ). Within the skills audit, students were asked to rate how important various research skills were to them. At the start of semester one, 77% of students rated knowing how GenAI operates as quite or extremely important, 79% thought prompt writing was quite or extremely important, 88% thought that understanding ethical use was quite or extremely important, and 90% reported that data integrity was quite or extremely important. No significant change in the value the students put on each aspect was recorded in the second iteration of the survey. Confidence in How GenAI Operates At the beginning of the semester, only 33% of students reported feeling "Quite" confident in their understanding of how GenAI operates, with 44% indicating "Limited" confidence and 23% reporting no confidence at all. By the end of the semester, confidence levels improved significantly, with 56% of students feeling "Quite" confident and only 19% reporting "Not" confident. This shift indicates a substantial increase in student understanding and confidence in operating GenAI tools. Confidence in Writing Prompts Initially, 52% of students expressed "Limited" confidence in their ability to write prompts effectively, and only 21% felt "Quite" confident. By the end of the semester, the proportion of students feeling "Quite" confident more than doubled to 48%, while those with "Limited" confidence decreased to 33%. This trend suggests that instructional efforts in prompt writing were successful in boosting student confidence. Confidence in the Ethical Use of AI Ethical considerations are a critical aspect of AI usage. At the start of the semester, only 42% of students reported being "Quite" confident in their understanding of the ethical use of AI, with 18% indicating no confidence at all. By the semester's end, "Quite" confidence rose to 67%, and the proportion of students with no confidence dropped to just 10%. This improvement reflects the effectiveness of the curriculum in enhancing students' awareness and confidence in ethical AI practices. Confidence in Data Protection Awareness Data protection awareness showed similar trends. Initially, 45% of students felt "Quite" confident in their understanding of data protection, while 43% reported "Limited" confidence and 12% had no confidence. By the end of the semester, those "Quite" confident increased to 71%, and those "Limited" decreased to 23%, with only 6% reporting no confidence. This increase demonstrates significant progress in students' understanding and confidence in data protection principles. Overall, the data indicates a clear positive trend in student confidence across all assessed areas. These results suggest that the semester's instruction effectively enhanced students' understanding and confidence in key AI-related competencies. Students’ voice During the third semester, fifteen one-on-one interviews were conducted to gauge the student cohort's perception of the changes made to the curriculum delivery in relation to their GenAI digital competence. Student researchers conducted all interviews to reduce data bias created by the perceived academic hierarchy between the MSc students and the interviewers. Each interview followed a specific question format as outlined in the supplementary data. The interview responses were transcribed and thematically coded, with three key themes being identified: 1) AI literacy and competence, 2) AI skill transferability, 3) Trust and Ethical Concerns. These themes are interdependent, forming a holistic experience of AI in academia. Increased utilisation leads to more learning and adaptation but also raises ethical concerns. AI's support is most effective when users have adapted to its use and trust it within ethical boundaries. The relationship between these themes is also cyclical. As students use AI and gain more experience, they refine their usage, adapt their learning, and continually reassess the ethical implications. This, in turn, influences how they continue to use AI and the support they look for from it. Theme 1 - AI literacy and competence Students shared how the structured educational modules have significantly contributed to their understanding of GenAI tools. The experience has not only increased their confidence in using GenAI but has also improved their ability to use GenAI effectively in a range of tasks. "I've enhanced my skills in asking the right questions to AI and using tools like ChatGPT and others effectively after the [semester 1] module." "My confidence in using AI has increased significantly thanks to the modules." Many students use GenAI to assist with tasks such as summarising research papers, paraphrasing text, and tailoring CVs. AI is used to clarify complex concepts that students may not fully understand from lectures. Non-native English speakers use AI to help with language-related tasks, such as grammar checks and translation. "It helps me with emails and checking my spelling because I am still learning English. It always translates for me or gives me synonyms, which makes it easier to write correctly." "Sometimes the slang in articles is difficult to understand, so this tool is helpful for us, especially for those improving their English." "I use AI to paraphrase and rephrase my writing, which helps me improve the quality of my work. It also provides feedback on grammar, which is crucial as I am still learning English." Theme 2 - AI skill transferability The learning process includes understanding the limitations of GenAI, developing prompting skills, and recognising the broader potential of AI in their academic and professional lives. "The modules have taught me how to use AI in a positive manner, whether for academic purposes or other tasks." "The module taught me tricks like reading papers, creating outlines, and even preparing for job searches. Without AI, starting something like exam prep or coursework would have been more difficult." Students view AI as a valuable assistant that provides essential support in various areas. This includes language assistance for non-native speakers, simplifying complex tasks, guidance in learning new concepts, and feedback to improve their work. AI is seen as a tool that makes complex tasks more manageable by providing quick answers, summaries, and guidance. As students learn more about AI, they begin to use it more effectively as a supportive tool, refining their prompts and better understanding how to get the assistance they need. "The module helped me understand the potential of AI beyond basic applications, such as writing essays. It showed me that AI could be a valuable tool if used correctly." "The module taught us how to use AI tools effectively, not just for coursework but also for presentations and job searches. It’s like having a mentor guiding you through different tasks." Theme 3 - Trust and Ethical Concerns students addressed concerns regarding GenAI's reliability and ethical use. Some students express doubts about the accuracy and reliability of AI, which affects their trust in using it for critical tasks, leading to a more informed and cautious approach to using these tools. "I use it sometimes but I'm not a big fan, you know. ChatGPT, for example. I don't really trust it because sometimes it gives wrong answers." Students raised concerns about the ethical implications of using AI, especially regarding plagiarism and the originality of work. This negatively impacted their confidence in using GenAI tools for their studies. "Using AI to produce essays or coursework is unethical. We should use it only for understanding or improving our own work." "We cannot put any personal info into AI because it could be discovered by others. I didn’t know it takes personal info and feeds it into the system." "I think I'm less confident in it after the [first semester] module because I didn’t realise things like plagiarism and data privacy." Students were also aware of the impact of AI on research integrity in the wider academic community, due to GenAI’s ability to generate “data” that hasn’t been empirically tested. "Nowadays, some people use AI to generate data or images for research without doing the actual experiments, which could undermine scientific integrity." Together, these themes describe a complex and dynamic interaction between the students and GenAI. GenAI’s role in learning is not just about utility but is deeply intertwined with ethical considerations, the learning process, and the level of support it GenAI provides. Discussion Integrating GenAI into the MSc skills development program demonstrates both the potential and the challenges of embedding such technologies into higher education. Our findings suggest that while GenAI tools can significantly enhance personalised learning experiences and student engagement, their use also necessitates careful consideration of ethical implications, particularly regarding academic integrity and the development of critical cognitive skills. Students in this study and elsewhere ((Ngo, 2023 ) generally have a favourable opinion of using ChatGPT in education, citing benefits like time savings and personalised tutoring. However, they also identify barriers such as the inability to assess the quality and reliability of sources and the inability to cite sources accurately. One of the primary advantages observed was the role of GenAI in providing personalised feedback and support to students, particularly those from non-native English-speaking backgrounds. GenAI tools can enhance communicative practices for non-native English speakers and support students' language learning experiences both in the classroom and outside of it. These tools address issues such as lack of motivation, anxiety, limited authentic communication opportunities, and lack of personalised feedback. Similar observations about the support that GenAI can give learners are also reported in further case studies (Bedford et al., 2024 ; Zadorozhnyy & Lai, 2023 ). Here, the students in question used GenAI tools as part of a personalised English language enhancement course, reporting that AI-assisted learning was an effective way to identify issues with their writing. Strategies include utilising GenAI as a mentor for enhancing sentence structure, grammar, and spelling, providing personalised learning strategies and resources, recommending relevant language learning applications tailored to specific skills, and offering follow-up questions for comprehension checks (Bedford et al., 2024 ; Zadorozhnyy & Lai, 2023 ). GenAI's ability to give students instant and personalised feedback on their writing and research tasks is, therefore, a promising development (Kasneci et al., 2023 ). However, concerns have been raised that prolonged use of GenAI tools may decrease higher-order thinking skills (Putra et al., 2023 ; Richardson & Clesham, 2021 ; Tenakwah et al., 2023 ). Tools like ChatGPT and ChatPDF were shown, here and elsewhere, to be instrumental in helping students navigate complex academic tasks (Elkhodr et al., 2023 ; Kasneci et al., 2023 ), from summarising research papers to refining their writing. This aligns with the constructivist approach to learning, where students actively construct knowledge through interactions with these AI tools (Zhou & Schofield, 2024 ). The ability of GenAI to offer tailored educational experiences is a significant advancement, making learning more accessible and engaging, especially for a diverse cohort (Guo et al., 2024 ; Popenici & Kerr, 2017 ). Such approaches allow students to receive tailored support while learning, enhancing engagement and educational outcomes aligning with individual learning trajectories (Bhutoria, 2022 ; Hooda et al., 2022 ). The increase in student confidence seen here across various GenAI-related competencies, such as prompt writing and ethical AI use, further underscores the value of integrating these technologies into the curriculum. The structured approach of embedding GenAI into core modules through the tip of the week, with “just-in-time” learning principles (Novak, 2011 ), allowed students to gradually build their skills, leading to a notable improvement in their self-assessed abilities by the end of the program. Just-in-time learning is rooted in delivering knowledge and skills precisely when needed (Welch, 2010 ). Embedding learning within relevant and meaningful situations, rather than teaching concepts in isolation, helps learners improve their understanding and application of concepts (Darling-Hammond & Snyder, 2000 ). This approach of applied learning is believed to enhance understanding, retention, and application of knowledge through real-life experiences and practical applications (Jach & Trolian, 2023 ). Students generally view AI as a helpful tool that supports them in various tasks. This support is most beneficial when users are familiar with using AI effectively and trust that it can provide reliable assistance. The more users understand AI, the more they can depend on it for effective support. Learning to use AI involves asking the right questions and interpreting AI's responses. Students may initially rely heavily on AI for support. As they become more aware of ethical concerns, they might adjust how they use AI to ensure they uphold academic integrity. As students become more proficient and confident in using AI, they learn and adapt their strategies for incorporating these tools into their academic work. This theme emphasises the learning process associated with mastering AI tools and adjusting to their continually evolving capabilities. Despite these benefits, the study also highlights significant concerns related to the ethical use of AI and the potential for academic dishonesty. The scepticism expressed by some students about the reliability of GenAI tools, coupled with the concerns about plagiarism and data privacy, indicates a need for ongoing education and clear guidelines on AI use in academic settings. The mixed responses regarding the appropriateness of AI in generating academic work suggest that while students appreciate the support AI can offer, they are also aware of the boundaries that must be maintained to ensure academic integrity. Clear guidelines for AI use in assessments can address these issues by setting out expectations and boundaries (Perkins, 2023 ). This was achieved here through structured process-driven assessments allowing students to log GenAI usage (Smith & Francis, 2024 ) and setting aside seminar time to detail institutional guidelines. Moorhouse et al. ( 2023 ), following their review of publicly available guidelines from the top 50 universities in the Times Higher Education World University Rankings 2023, noted a move to a redesign of assessments to effectively incorporate GenAI tools designing tasks that emphasise critical thinking and creativity (Bobula, 2024 ; Chan, 2023 ; Moorhouse et al., 2023 ). Assessment methods include open-ended questions (Meir et al., 2024 ), project-based assessments, and working-world problem-solving tasks (Petrovska et al., 2024 ) that encourage original thought and application of knowledge, thereby reducing the potential for AI misuse (Meir et al., 2024 ). As GenAI technologies continue to evolve, there is a risk that students might bypass critical thinking processes, leading to a superficial understanding of the subject matter. The potential for over-reliance on AI tools, particularly in producing assessments, raises questions about the long-term impact on students' cognitive and creative skills (Richardson & Clesham, 2021 ; Tenakwah et al., 2023 ). GenAI, use in learning, should complement rather than replace the essential human elements of learning. The success of this program can be attributed to the thoughtful integration of GenAI into the curriculum through experience mapping (Beard, 2022 ) and the deliberate design of assessments (Smith & Francis, 2024 ). By aligning AI tools with core learning objectives and providing students with opportunities to critically engage with AI outputs, the program helped students develop a nuanced understanding of both the capabilities and limitations of GenAI. However, the study also indicates that the effectiveness of such integrations depends heavily on the design and delivery of the curriculum. The cyclical relationship observed between students’ increasing experience with GenAI and their evolving perceptions and ethical considerations highlights the need for continuous adaptation and refinement of teaching strategies. This iterative process is crucial in ensuring that GenAI tools enhance, rather than detract from, the educational experience. Implications for Future Research and Practice This study provides valuable insights into the practical implementation of GenAI in higher education and opens several avenues for future research. The long-term effects of GenAI on student learning outcomes, particularly in terms of developing higher-order thinking skills, need to be explored. Additionally, further research is needed to evaluate the impact of AI on different student demographics, especially those who may be disproportionately affected by AI-related biases. Educational institutions must prioritise developing comprehensive AI literacy programs for students and educators. These programs should foster a critical understanding of GenAI, encourage ethical use, and equip students with the skills necessary to navigate an increasingly AI-driven world. GenAI is being integrated into academic activities, serving as both a learning aid and a productivity tool. However, its use raises important questions about academic integrity and the potential for over-reliance. While AI tools are beneficial, students are rightly cautious about relying on them too heavily. This scepticism highlights the need for continued education on the ethical use of GenAI and the importance of critical thinking when interpreting AI-generated content. Structured learning is crucial in helping students gain confidence in using AI, which suggests that formal education on AI can significantly impact how effectively students utilise these tools. Abbreviations AI artificial intelligence GenAI generative artificial intelligence LLMs large language models Declarations GenAI declaration GTP4 (OpenAI) was used to create a broad paper outline and suggest content areas for discussion. The code used to generate the figures was drafted in GTP4 using dummy data before being edited and implemented in R, ensuring data integrity. Consensus was used to find and summarise relevant research articles for the introduction and discussion. Text editing was completed using Grammarly. Availability of data and materials The datasets generated and/or analysed during the current study are not publicly available due to the confidential nature of the transcripts generated, but they are available from the corresponding author upon reasonable request. Competing interests None Funding This work was funded by a teaching and learning development grant from the College of Health, Wellbeing and Life Sciences (HWLS), Sheffield Hallam University, Sheffield, S1 1WB Authors' contributions DPS conceived and designed the project, and DS, SM, CO, and CB acquired the data. All authors were involved in analysing and interpreting the data, and DPS, MML, and NJF wrote the manuscript. Acknowledgements We would like to acknowledge Dr Marjory Da Costa Abreu for her collaboration and input into the project. For the purpose of open access, the author has applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version of this paper arising from this submission. References Ali, O., Murray, P. A., Momin, M., Dwivedi, Y. K., & Malik, T. (2024). The effects of artificial intelligence applications in educational settings: Challenges and strategies. Technological Forecasting and Social Change, 199 , 123076. Beard, C. (2022). Experiential learning design: Theoretical foundations and effective principles . Routledge. Bearman, M., & Luckin, R. (2020). Preparing university assessment for a world with AI: Tasks for human intelligence. In M. Bearman, P. Dawson, R. Ajjawi, J. Tai & D. Boud (Eds.), Re-imagining university assessment in a digital world (pp. 49–63). Springer International Publishing. https://doi.org/10.1007/978-3-030-41956-1_5 Bedford, J., Kim, M., & Qin, J. C. (2024). Confidence enhancer, learning equalizer, and pedagogical ally. Using generative AI effectively in higher education: Sustainable and ethical practices for learning, teaching and assessment (1st ed., pp. 33). Taylor & Francis. https://doi.org/10.4324/9781003482918-17 Bhutoria, A. (2022). Personalized education and artificial intelligence in the united states, china, and india: A systematic review using a human-in-the-loop model. Computers and Education: Artificial Intelligence, 3 , 100068. https://doi.org/10.1016/j.caeai.2022.100068 Bobula, M. (2024). Generative artificial intelligence (AI) in higher education: A comprehensive review of challenges, opportunities, and implications. Journal of Learning Development in Higher Education, (30)https://doi.org/10.47408/jldhe.vi30.1137 Brew, M., Taylor, S., Lam, R., Havemann, L., & Nerantzi, C. (2023). Towards developing AI literacy: Three student provocations on AI in higher education. Asian Journal of Distance Education, 18 (2), 1–11. https://doi.org/10.5281/zenodo.8032387 Chan, C. K. Y. (2023). A comprehensive AI policy education framework for university teaching and learning. International Journal of Educational Technology in Higher Education, 20 (1), 38. Chan, C. K. Y., & Colloton, T. (2024). Generative AI in higher education: The ChatGPT effect . Taylor & Francis. https://doi.org/10.4324/9781003459026 Crowther, M., Lim, W., & Crowther, M. A. (2010). Systematic review and meta-analysis methodology. Blood, the Journal of the American Society of Hematology, 116 (17), 3140–3146. Darling-Hammond, L., & Snyder, J. (2000). Authentic assessment of teaching in context. Teaching and Teacher Education, 16 (5-6), 523–545. Elkhatat, A. M., Elsaid, K., & Almeer, S. (2023). Evaluating the efficacy of AI content detection tools in differentiating between human and AI-generated text. International Journal for Educational Integrity, 19 (1), 17. Elkhodr, M., Gide, E., Wu, R., & Darwish, O. (2023). ICT students’ perceptions towards ChatGPT: An experimental reflective lab analysis. STEM Education, 3 (2), 70–88. Fosnot, C. T. (2013). Constructivism: Theory, perspectives, and practice . Teachers College Press. Guo, H., Yi, W., & Liu, K. (2024). Enhancing constructivist learning: The role of generative AI in personalised learning experiences. Proceedings of the 26th International Conference on Enterprise Information Systems (ICEIS 2024), 1 , 767–770. https://doi.org/10.5220/0012688700003690 Hooda, M., Rana, C., Dahiya, O., Rizwan, A., & Hossain, M. S. (2022). Artificial intelligence for assessment and feedback to enhance student success in higher education. Mathematical Problems in Engineering, 2022 (1), 5215722. Huxley-Binns, R., Lawrence, J., & Scott, G. (2023). Competence-based HE: Future proofing curricula. Integrative curricula: A multi-dimensional approach to pedagogy (pp. 131–147). Emerald Publishing Limited. Jach, E., & Trolian, T. (2023). Supporting college student success through applied learning: Considering associations with average college grades, graduation in four years, and degree aspirations. Journal of Postsecondary Student Success, 2 (3), 75–97. Jeltova, I., Birney, D., Fredine, N., Jarvin, L., Sternberg, R. J., & Grigorenko, E. L. (2007). Dynamic assessment as a process-oriented assessment in educational settings. Advances in Speech Language Pathology, 9 (4), 273–285. Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., & Hüllermeier, E. (2023). ChatGPT for good? on opportunities and challenges of large language models for education. Learning and Individual Differences, 103 , 102274. Lacey, M. M., & Smith, D. P. (2023). Teaching and assessment of the future today: Higher education and AI. Microbiology Australia, 44 (3), 124–126. https://doi.org/10.1071/MA23036 Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., & Zou, J. (2023). GPT detectors are biased against non-native english writers. Patterns, 4 (7), 100779. https://doi.org/10.1016/j.patter.2023.100779 Lodge, J. M., de Barba, P., & Broadbent, J. (2023). Learning with generative artificial intelligence within a network of co-regulation. Journal of University Teaching and Learning Practice, 20 (7), 1–10. Meir, E., Pope, D., Abraham, J. K., Kim, K. J., Maruca, S., & Palacio, J. (2024). Designing activities to teach higher-order skills: How feedback and constraint affect learning of experimental design. CBE—Life Sciences Education, 23 (1), ar1. Moorhouse, B. L., Yeo, M. A., & Wan, Y. (2023). Generative AI tools and assessment: Guidelines of the world's top-ranking universities. Computers and Education Open, 5 , 100151. Ngo, T. T. A. (2023). The perception by university students of the use of ChatGPT in education. International Journal of Emerging Technologies in Learning (Online), 18 (17), 4. Noroozi, O., Soleimani, S., Farrokhnia, M., & Banihashem, S. K. (2024). Generative AI in education: Pedagogical, theoretical, and methodological perspectives. International Journal of Technology in Education, 7 (3), 373–385. Novak, G. M. (2011). Just‐in‐time teaching. New Directions for Teaching and Learning, 2011 (128), 63–73. Palmer, E., Lee, D., Arnold, M., Lekkas, D., Plastow, K., Ploeckl, F., Srivastav, A., & Strelan, P. (2023). Findings from a survey looking at attitudes towards AI and its use in teaching, learning and research. ASCILITE Publications, https://doi.org/10.14742/apubs.2023.537 Perkins, M. (2023). Academic integrity considerations of AI large language models in the post-pandemic era: ChatGPT and beyond. Journal of University Teaching & Learning Practice, 20 (2)https://doi.org/10.53761/1.20.02.07 Petrovska, O., Clift, L., Moller, F., & Pearsall, R. (2024). Incorporating generative AI into software development education. Paper presented at the Proceedings of the 8th Conference on Computing Education Practice, 37–40. Popenici, S. A., & Kerr, S. (2017). Exploring the impact of artificial intelligence on teaching and learning in higher education. Research and Practice in Technology Enhanced Learning, 12 (1), 22. Putra, F. W., Rangka, I. B., Aminah, S., & Aditama, M. H. (2023). ChatGPT in the higher education environment: Perspectives from the theory of high order thinking skills. Journal of Public Health, 45 (4), e840–e841. Richardson, M., & Clesham, R. (2021). Rise of the machines? the evolving role of artificial intelligence (AI) technologies in high stakes assessment. London Review of Education, 19 (1), 1–13. Rodriguez, J. G., Hunter, K. H., Scharlott, L. J., & Becker, N. M. (2020). A review of research on process oriented guided inquiry learning: Implications for research and practice. Journal of Chemical Education, 97 (10), 3506–3520. Romova, Z., & Andrew, M. (2011). Teaching and assessing academic writing via the portfolio: Benefits for learners of english as an additional language. Assessing Writing, 16 (2), 111–122. Rudolph, J., Tan, S., & Tan, S. (2023). ChatGPT: Bullshit spewer or the end of traditional assessments in higher education? Journal of Applied Learning and Teaching, 6 (1), 342–363. Sallam, M. (2023). ChatGPT utility in healthcare education, research, and practice: Systematic review on the promising perspectives and valid concerns. Healthcare, 11 (6), 887. https://doi.org/10.3390/healthcare11060887 Schreiber, N., Theyßen, H., & Schecker, H. (2016). Process-oriented and product-oriented assessment of experimental skills in physics: A comparison. Paper presented at the Insights from Research in Science Teaching and Learning: Selected Papers from the ESERA 2013 Conference, 29–43. Smith, D., & Francis, N. (2024). Process not product in the written assessment. Using generative AI effectively in higher education (1st ed., pp. 115–126). Routledge. https://doi.org/10.4324/9781003482918-17 Sok, S., & Heng, K. (2024). Opportunities, challenges, and strategies for using ChatGPT in higher education: A literature review. Journal of Digital Educational Technology, 4 (1), ep2401. https://doi.org/10.30935/jdet/14027 Sotiriadou, P., Logan, D., Daly, A., & Guest, R. (2020). The role of authentic assessment to preserve academic integrity and promote skill development and employability. Studies in Higher Education, 45 (11), 2132–2148. https://doi.org/10.1080/03075079.2019.1582015 Tenakwah, E. S., Boadu, G., Tenakwah, E. J., Parzakonis, M., Brady, M., Kansiime, P., Said, S., Ayilu, R., Radavoi, C., & Berman, A. (2023). Generative AI and higher education assessments: A competency-based analysis. Res.Sq, https://doi.org/10.21203/rs.3.rs-2968456/v2 Welch, R. (2010). How just in time learning should become the norm! Paper presented at the 2010 Annual Conference & Exposition, 15.649. 1–15.649. 12. Zadorozhnyy, A., & Lai, W. Y. W. (2023). ChatGPT and L2 written communication: A game-changer or just another tool? Languages, 9 (1), 5. Zhou, X., & Schofield, L. (2024). Using social learning theories to explore the role of generative artificial intelligence (AI) in collaborative learning. Journal of Learning Development in Higher Education, (30)https://doi.org/10.47408/jldhe.vi30.1031 Additional Declarations The authors declare no competing interests. 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-5204546","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":362340777,"identity":"0cd2bb55-c63d-455d-bfa6-6d20e84200a4","order_by":0,"name":"David P. Smith","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA20lEQVRIiWNgGAWjYDCCA2BSgsGAvYENKsSGUzGKFgkDngOkaWGQMJBIIFIL3wHeA4w/cyzqzCWfP3vMw2AnzyCRloBXi+QBvgRm3m0SEpazc8yNeRiSDRsk0g7g1WJwgMeAmRGoxeB2Dps0DwNzAoNEegNBLYw/QVpuHn8G1FJPnBYGkMMMbjCYAbUcBmoh4DDJw3wJh4FaJDecyTGTnGNw3LCN51kCXi18x3sPPvy5rY7f4PjxZxJvKqrl+dnTDPBqYWDmgUUN2J2EIxIIeAgrGQWjYBSMghEOAHzOOraGypAJAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0001-5177-8574","institution":"Sheffield Hallam University","correspondingAuthor":true,"prefix":"","firstName":"David","middleName":"P.","lastName":"Smith","suffix":""},{"id":362340778,"identity":"edfd574c-e51c-48cb-a6e7-e20086d0d370","order_by":1,"name":"Dami Sokoya","email":"","orcid":"","institution":"Sheffield Hallam University","correspondingAuthor":false,"prefix":"","firstName":"Dami","middleName":"","lastName":"Sokoya","suffix":""},{"id":362340779,"identity":"43c5bd1f-fa32-4c7f-aa0e-4d6154816334","order_by":2,"name":"Skye Moore","email":"","orcid":"","institution":"Sheffield Hallam University","correspondingAuthor":false,"prefix":"","firstName":"Skye","middleName":"","lastName":"Moore","suffix":""},{"id":362340780,"identity":"923de681-a61f-4855-b8d6-c4b640efc0da","order_by":3,"name":"Chinenya Okonkwo","email":"","orcid":"https://orcid.org/0009-0006-5139-0477","institution":"Sheffield Hallam University","correspondingAuthor":false,"prefix":"","firstName":"Chinenya","middleName":"","lastName":"Okonkwo","suffix":""},{"id":362340781,"identity":"707328c3-54ec-42d5-a980-bff0ed73b48b","order_by":4,"name":"Charlotte Boyd","email":"","orcid":"https://orcid.org/0009-0002-5210-0647","institution":"Sheffield Hallam University","correspondingAuthor":false,"prefix":"","firstName":"Charlotte","middleName":"","lastName":"Boyd","suffix":""},{"id":362340782,"identity":"71750500-f2a4-4dd0-b106-8c771aecf52d","order_by":5,"name":"Melissa M. Lacey","email":"","orcid":"https://orcid.org/0000-0003-0997-0217","institution":"Sheffield Hallam University","correspondingAuthor":false,"prefix":"","firstName":"Melissa","middleName":"M.","lastName":"Lacey","suffix":""},{"id":362340783,"identity":"f54b6a77-8f36-4377-95b8-99139008dea7","order_by":6,"name":"Nigel J. Francis","email":"","orcid":"https://orcid.org/0000-0002-4706-4795","institution":"Cardiff University","correspondingAuthor":false,"prefix":"","firstName":"Nigel","middleName":"J.","lastName":"Francis","suffix":""}],"badges":[],"createdAt":"2024-10-04 14:16:27","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-5204546/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5204546/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":66169968,"identity":"c58bab93-5768-449a-b010-570c630b9f1d","added_by":"auto","created_at":"2024-10-08 10:31:07","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":268883,"visible":true,"origin":"","legend":"\u003cp\u003eSimplified Experience Map of the MSc skills development program. Each Semester is presented as a zone on the map. Assessments are represented as open circles. The delivery and content of GenAI development content are highlighted on the green line. Semester one: explanation of how LLM generate responses, ethical and data integrity considerations; Semester two: GenAI usages presented in the context of understanding content, reflective practice and employability skills. The timing of evaluation points is highlighted.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-5204546/v1/496fcee33d817720a64b84b7.png"},{"id":66169972,"identity":"15ba8ee2-8cc3-4b28-95c6-df575d32cabc","added_by":"auto","created_at":"2024-10-08 10:31:07","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":177203,"visible":true,"origin":"","legend":"\u003cp\u003eThe number of students who use GenAI tools was determined by asking the question, “How often do you use GenAI tools?” n=101 responses.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-5204546/v1/b25d880d90d48acab8ac73f2.png"},{"id":66169969,"identity":"ecc1a1cb-36c4-4bbc-bc74-c3f667e16277","added_by":"auto","created_at":"2024-10-08 10:31:07","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":103000,"visible":true,"origin":"","legend":"\u003cp\u003ePercentage of respondents' perceptions regarding the usefulness of AI in four academic tasks: writing assessments, helping to understand subject content, editing written content, and finding and understanding research articles. The chart shows the distribution of responses across five categories: Strongly Disagree, Disagree, Neutral, Agree, and Strongly Agree.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-5204546/v1/ca6238062a1d32f72e14a15d.png"},{"id":66169970,"identity":"3ed34e96-a499-440f-a039-0d3b198f7514","added_by":"auto","created_at":"2024-10-08 10:31:07","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":123980,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of student confidence levels in AI-related skills between the start and end of the semester. The figure presents the percentage distribution of confidence levels across four key skills: understanding how generative AI (GenAI) operates, writing prompts, ethical use of AI, and data protection awareness. Each skill is measured at two points in time, \"Start\" (beginning of the semester) and \"End\" (end of the semester). Confidence levels are categorised into five levels: \"Not,\" \"Limited,\" \"Somewhat,\" \"Quite,\" and \"Extremely.\" Mann–Whitney, U-tests between groups were used. Statistical significance is indicated as P \u0026lt; 0.05, P \u0026lt; 0.01.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-5204546/v1/5f53a6afa503a774d40cc84b.png"},{"id":66170931,"identity":"e2af6229-a1bf-4604-84b4-b7f4dc301eb1","added_by":"auto","created_at":"2024-10-08 10:39:07","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1226316,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5204546/v1/b1348c6c-bc6e-4103-8bce-1c36f5c22a2e.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eEmbedding Generative AI as a digital capability into a year-long MSc skills program\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eGenerative Artificial Intelligence (GenAI) has quickly become a transformative factor in higher education, influencing assessment practices and pedagogical approaches to learning and teaching. Large Language Models (LLMs) offer unprecedented capabilities for personalised learning journeys (Bearman \u0026amp; Luckin, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Hooda et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). GenAI can deliver substantial efficiency gains and enhanced educational outcomes through personalised learning pathways when considered an enabler. However, the disruptive nature of the technology and its emergence in higher education has raised concerns about academic integrity, the authenticity of student work, and the potential reduction in student development of essential cognitive and creative skills (Richardson \u0026amp; Clesham, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Tenakwah et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eLarge Language Models (LLMs) are trained on large collections of text data to capture as much human language as possible (Rudolph et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). When asked questions, LLMs predict which words to respond with based on the training set. The GenAI recalls prior interactions during conversations to inform its next prediction, allowing for conversational interactions. Supporters of GenAI argue that it can enhance efficiency, productivity, and information access; however, critics are concerned that GenAI could lead to declining academic standards (Sallam, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Due to the training sets used in creating LLMs, there is an inherent bias stemming from prejudices in the source material that reflect historical or social inequities. It is crucial to ensure that GenAI enhances human skills and understanding rather than replacing them (Bearman \u0026amp; Luckin, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Bobula, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This approach involves using clear frameworks and constantly evaluating and updating assessment practices to leverage GenAI's capabilities while maintaining the educational experience's integrity.\u003c/p\u003e \u003cp\u003eGenAI has shown its rapidly evolving capabilities in producing realistic human-like text, convincing images, videos, and other media. Many evaluation methods rely on the capability to generate new content (the evaluation) based on a prompt (the evaluation brief) using existing knowledge (the training dataset). The emergence of AI-generated content presents clear challenges for preserving academic honesty in this context. Richardson and Clesham (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) point out the difficulties in identifying AI-generated submissions, which can undermine the credibility of student work. Concerns about academic honesty in the context of artificial intelligence (AI) are supported by data showing challenges distinguishing between student-generated and AI-generated work (Elkhatat et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). There are also significant concerns that AI detectors unfairly put neurodiverse and non-native English language speakers at a disadvantage (Liang et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The use of AI in education gives rise to various ethical concerns. It is crucial that AI systems are developed and implemented in a way that prevents the reinforcement of existing biases and promotes equal educational opportunities. Addressing these risks entails the establishment of clear guidelines for the use of AI in assessments (Perkins, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe effective integration of AI tools into learning environments requires strategic planning and a commitment to pedagogical principles. GenAI tools have the potential to significantly enhance educational outcomes by providing personalised feedback, facilitating language learning, and supporting various research methodologies. Additionally, they can increase learner engagement and motivation while maintaining a strong ethical framework that focuses on privacy, bias, and accuracy. Understanding the impact of GenAI on education necessitates a comprehensive theoretical approach (Noroozi et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Constructivism proposes that learning is an active process wherein learners build their understanding of the world through experiences and interactions with the environment (Fosnot, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). The integration of GenAI into this framework echoes social constructivism, where learning is viewed as an active process in which knowledge is constructed through social interactions (Zhou \u0026amp; Schofield, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). GenAI tools can facilitate collaborative learning and provide personalised feedback to enhance student engagement. Much of the power of GenAI lies in the creation of personalised learning environments through tools and systems that tailor educational experiences to individual student needs. These tools can dynamically adjust content delivery and assessment methods based on continuous data analysis, enhancing educational outcomes by aligning with individual learning trajectories. Competency-based learning assesses students\u0026rsquo; ability to apply concepts in real-world scenarios rather than through traditional memory-based tasks (Huxley-Binns et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Behaviourist principles, particularly reinforcement and feedback, are central to competency-based education. In this framework, learning is seen as a change in behaviour resulting from acquiring knowledge or skills, and students are assessed based on observable outcomes or competencies. This approach assesses students' ability to apply concepts in real-world scenarios. GenAI can facilitate competency-based learning by providing detailed feedback on performance and identifying areas for improvement, aiding learners in mastering specific competencies.\u003c/p\u003e \u003cp\u003eIn 2023, Moorhouse et al. (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) reviewed guidelines from the top 50 universities in the Times Higher Education World University Rankings 2023. The review focused on the content and advice provided for instructors about using GenAI in assessments. The review encouraged redesigning assessments to incorporate GenAI tools effectively. This includes designing tasks with GenAI, emphasising critical thinking and creativity, and using in-class assessments to prevent potential misuse. Emphasising the learning process over final outputs and supporting staged assessments that allow for feedback and development were seen as good practices.\u003c/p\u003e \u003cp\u003eAI's capabilities require innovative approaches to assessment design with a shift to creating assessments that demand critical thinking and a deeper understanding of the subject area (Bobula, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Chan \u0026amp; Colloton, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Assessment methods include open-ended questions (Meir et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), project-based assessments (Petrovska et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), and real-world problem-solving tasks that encourage original thought and application of knowledge. There is a need for a significant shift in thinking about assessment design and the iterative creation of the work to be submitted. Project-based learning addresses this need, where students must engage in a process over time, such as research projects, case studies, or design tasks, which requires original thinking and continuous instructor feedback (Petrovska et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Developing assessments that explicitly require students to use AI tools for certain parts of the task, followed by a critical evaluation of the AI\u0026rsquo;s output, can help students develop critical thinking skills about AI and its capabilities (Sok \u0026amp; Heng, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). By using a product-oriented rather than a process-driven assessment model, it is more likely that even if students use GenAI to complete assessments, they will have to engage with, refine, and update the outputs of tools and evidence ethical usage of the AI package selected (Smith \u0026amp; Francis, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSurveys of educators and students reveal diverse attitudes towards AI in education (Chan \u0026amp; Colloton, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Lacey \u0026amp; Smith, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Palmer et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The analysis indicates a general recognition of AI's benefits in supplementing learning, assessment efficiency and feedback. However, there is also significant apprehension regarding the potential for academic dishonesty and its impact on learning processes. Many educators see the potential of AI in teaching but have concerns about ethical issues and academic integrity (Lodge et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The use of generative AI in education challenges traditional practices and requires new approaches to foster creativity alongside the risk and the risk of undermining the authenticity of student work (Ali et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Educators are also facing complexities and constraints in adapting these new digital tools into their practices and feel the need for AI literacy and support in this new digital competency. Student responses to GenAI in assessments vary. Chan \u0026amp; Colloton (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) found that some students appreciate the personalised feedback and learning support that GenAI offers, while others worry about how it might affect their creativity and originality. These concerns highlight the importance of clear institutional policies and guidelines, as well as the necessity of educating students about ethical AI usage (Smith \u0026amp; Francis, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Although students generally trust AI for tasks like grammar correction, they still prefer human educators for assessment feedback (Palmer et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Smith \u0026amp; Francis, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). When comparing attitudes across different assessment situations, it is clear that tailored approaches to AI integration are needed to address specific worries and contexts. Recommendations include establishing comprehensive guidelines for AI use, ensuring that AI complements rather than replaces traditional teaching methods, and emphasising responsible and ethical implementation. By understanding and addressing these diverse perspectives, educational institutions can better navigate the challenges and opportunities brought by AI in assessment practices.\u003c/p\u003e \u003cp\u003eWhile GenAI offers tremendous potential to enhance personalised learning experiences and educational outcomes, its integration into higher education must carefully consider academic integrity, ethical concerns, and ensuring that it complements, rather than replaces, essential human cognitive and creative skills. In this study, experience mapping, a concept developed by Colin Beard as part of his broader framework on experiential learning (Beard, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), was used to embed GenAI into a series of three core research skills modules. The process involves designing, analysing, and refining learning experiences to maximise their impact on learners, delivering and building core skills and competencies through structured delivery. Incorporating AI into module assessments in this study involved a shift towards emphasising the creation process rather than the product (Smith \u0026amp; Francis, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), especially with written articles (Rudolph et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Assessing the process over the product has several advantages as it integrates tool usage as part of the overall output and is independent of the tool used, giving the assessment longevity (Romova \u0026amp; Andrew, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Additionally, there are benefits to addressing academic integrity by structuring assessment tasks that contribute towards the final article (Sotiriadou et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) emphasises the learning journey and the steps taken rather than just evaluating the final product (Jeltova et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Rodriguez et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Rudolph et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Schreiber et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHere, we evaluate a year-long study in which GenAI was embedded into the skills development program as part of an MSc program. We aimed to determine if student understanding and confidence in the use of GenAI could be delivered and enhanced in a large cohort teaching environment. To achieve this, GenAI was integrated into core modules delivered each semester over a two-semester research skills program. Students reported an increase in confidence around GenAI. Thematic analysis of open-text response questions and interview transcripts describes a cyclical relationship between themes. As students use GenAI and gain more experience, they refine their usage, adapt their learning, and continually reassess the ethical implications. This, in turn, influences how they continue to use GenAI and the support they seek from it.\u003c/p\u003e \n"},{"header":"Methods","content":"\u003ch3\u003eCurriculum delivery\u003c/h3\u003e\n\u003cp\u003eThe experience mapping process utilised here involved designing, analysing, and refining learning experiences to maximise their impact on learners and was used to create a visual representation of the learning journey (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). This map helps educators and students see the sequence of experiences, identifying critical touchpoints and was used to plan how different elements of GenAI skills development could be incorporated.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDuring the first semester, points were identified as the need to understand GenAI's workings, the ethical and appropriate use of the tools, data integrity considerations and the institutional policies around academic conduct. These moments were crucial for understanding how AI produces its responses and reflected on the use of GenAI during their studies. During semester two, prompt design and example GenAI tools were covered that addressed specific needs, such as aiding in understanding, scaffold reflection, feeding forward on assessment tasks, or acting as a learning guide.\u003c/p\u003e \u003cp\u003eDuring semester one, an introduction to GenAI and how it operates alongside the ethical and appropriate use of GenAI was delivered. During semester two, prompt engineering, examples and different tools were embedded into the taught delivery. Tool usage supported by video content and prompt design were appropriate to the topic and were delivered using the just-in-time principle, with example prompts supporting the activities of the week. These were presented to the students as the \u0026ldquo;GenAI Tip of the week,\u0026rdquo; delivered during seminars and supported by prompt libraries and video support via the virtual learning environment.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003ePrompt\u003c/strong\u003e \u003cp\u003eBreak down [topic you\u0026rsquo;d like to understand] into smaller, easier-to-understand parts. Use analogies and real-life examples to simplify and make the concept more relatable.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003ePrompt\u003c/strong\u003e \u003cp\u003eHow would you verify the information in this conversation?\u003c/p\u003e \u003c/p\u003e \u003cp\u003e*Example prompt used to help understand more complex topics.\u003c/p\u003e \u003cp\u003eGenAI prompts were also presented to facilitate the reflection process by analysing learner inputs and providing automated yet personalised feedback.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003ePrompt\u003c/strong\u003e \u003cp\u003eAs an MSc student enrolled on a Bioscience or Chemistry course in the UK, you require an action plan and tips to complete your next assignment. You have received feedback on your lab bookkeeping assessment which includes some example areas that need improvement. Please review the feedback and provide suggestions for enhancing the laboratory report assessment.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e*Example prompt used to gain feedforward advice following the initial assessment.\u003c/p\u003e \u003cp\u003eReal-time Adaptation of Learning Paths was also presented, enabling real-time adaptation based on ongoing learner performance. GenAI prompts tools were presented that modify the learning path instantaneously, offering more challenges or support as needed.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003ePrompt\u003c/strong\u003e \u003cp\u003eYou are an [MSc Pharmaceutical Bioscience Student]. You are about to prepare a two-page CV to include with your application for a [PhD position]. Set out structure and ideas for content for an impactful CV.\u0026rdquo;\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003ePrompt\u003c/strong\u003e \u003cp\u003eIn this conversation, you will take on the role of an interviewer. You are looking to hire an intern for one year. The pharmaceutical company is looking for a bench scientist. Ask questions that would be suitable for this role. You will ask the question, and I will then give you my answer. You will then give feedback.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eGenAI tools and prompts were also presented to help learners who struggled with a particular concept particularly understanding research articles or choosing the correct statistical method for a given situation. Prompts were provided to generating additional examples, exercises, or explanations to reinforce understanding.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003ePrompt\u003c/strong\u003e \u003cp\u003eYour role in this conversation is to act as a guide helping a researcher to choose which statistical test to use when analysing their data. Your conversation will be based on working through a statistical decision tree. You will ask the researcher questions to guide them to the most appropriate statistical test. In your responses, give an explanation of the terms used below the main text. Use examples to help them understand. Your first question will be about the number of groups used in the study and ask for some background information. Are you ready?\u003c/p\u003e \u003c/p\u003e \u003cp\u003eAssessment design incorporated GenAI prompts, ensuring the rationale for tool utilisation and critiquing source validity (Brew et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Assessments were fully integrated into the taught context to help students develop these competencies and demonstrate learning outcomes. Seminars, and tutorials were conducted to foster the development of these competencies, including the ethical and proper use of GenAI, such as logging search strategies and keywords and critically appraising content validity (Crowther et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). GenAI was integrated into the assessment design by documenting prompts and justifying how the generated text contributed to the final article. Furthermore, templated guides were developed to structure this process, providing students with direction in documenting their progress towards the final product. A complete theoretical underpinning and assessment design strategy alongside assessment rubrics can be found in (Smith \u0026amp; Francis, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eA questionnaire was undertaken at the start of semester 1 to inform curriculum content based on students\u0026rsquo; current understanding and perceptions of AI. Students\u0026rsquo; self-reporting skills in a skills audit at the start of semester 1 and the end of semester 2 determined the impact of the teaching and learning within the curriculum. To allow for a more detailed understanding of the student experience, interviews with students were undertaken in semester 3.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003eThe student participants were MSc cohorts on a range of Biosciences and Chemistry master\u0026rsquo;s degree programmes comprising Analytical Chemistry and Pharmaceutical Analysis. Biomolecular Science, Biotechnology and Pharmacology, Microbiology Biology and Cancer Biology. The cohort is drawn from various nationalities, primarily Nigeria, India, Pakistan and the United Kingdom. In all degree programmes, students participate in core skills modules that run for three semesters, with weekly tutorials and seminars. Seminar sessions are co-taught with the full cohort and are split into tutorial groups of 25\u0026ndash;30 students. Laboratory sessions are a mandatory component of the degree programmes linked directly to assessment. From a total cohort of 180 students, 156 opted into the study (87% opt-in rate).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEthics\u003c/h3\u003e\n\u003cp\u003e Ethics for this study were acquired through the College of Health, Wellbeing, and Life Sciences ethics committee following the Sheffield Hallam University Research Ethics Policy (ER61054725). Ethical approval was given as no identifiable, confidential, or controversial information would be collected. No gender, age, educational experience or other demographic factors were requested or considered in the analysis. Participation in the study was optional.\u003c/p\u003e \u003cp\u003eThe following statement was displayed before in-class polling took place.\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eWe would like to evaluate your responses so that we can further improve our skills development program over the coming years specifically around AI and digital skills. Responses to the in-class poll are fully anonymised; no personally identifiable information will be collected. If you do not wish to take part please note your responses for your own records\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eBefore data collection for the skills audit, students were read and given a copy of the following statement, which served as a means of consent.\u003c/p\u003e \u003cp\u003e\u0026ldquo;We would like to evaluate your responses so that we can further improve our skills development program over the coming years specifically around AI and digital skills. To do this we will collect your responses in semester 1 and semester 2 and ask the question how have your skills developed. We will use your SHU e-mail to track your responses and the details will only be available to the study organised by Professor David Smith.\u003c/p\u003e \u003cp\u003eTaking part is entirely voluntary. Not taking part will not affect your studies or assessments. Input your student number if you consent to your anonymised responses being evaluated.\u0026rdquo;\u003c/p\u003e \u003cp\u003eStudents were recruited for the interviews via e-mails from the virtual learning environment and posters displayed in the teaching laboratories. As an incentive, a ten-pound gift card was given. Participants again gave consent by completing an interview consent form after reading the study information sheet (Supplementary Information).\u003c/p\u003e\n\u003ch3\u003eEvaluation Instruments\u003c/h3\u003e\n\u003cp\u003e \u003cstrong\u003eQuestionnaires\u003c/strong\u003e \u003cp\u003eThe student questionnaires (Supplementary Information) were collected, and an independent researcher indexed and transposed the details and comments into Excel. Students' names in the study were replaced with an identifier to prevent linking the responses to an individual and bias by the investigator. Responses to Likert scale questions were converted to numbers from 1 to 5 for statistical analysis, with 1 representing the most negative response to the question (e.g. strongly disagree) and 5 the most positive (e.g. strongly agree). The ordinal nature of the questionnaire data meant that parametric statistics were inappropriate for analysis, so nonparametric tests were used throughout. Observations were independent, with no individuals belonging to more than one study group. For Mann\u0026ndash;Whitney, U-tests between groups were used as appropriate. Statistical significance is indicated as P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, P\u0026thinsp;\u0026lt;\u0026thinsp;0.01.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eSkills Audit\u003c/strong\u003e \u003cp\u003eSkills audits (supplementary information) were conducted early in semester one and at the end of semester two. Data was collected via an online questionnaire during a large group-taught session. Students were asked to evaluate the importance and confidence of a range of research skills, including GenAI, on a Linkert scale of 1\u0026thinsp;=\u0026thinsp;not important/confident, 2\u0026thinsp;=\u0026thinsp;Limited importance/confident, 3\u0026thinsp;=\u0026thinsp;somewhat important/confident, 4\u0026thinsp;=\u0026thinsp;quite important/confident, and 5\u0026thinsp;=\u0026thinsp;extremely important/confident. A copy of the responses was e-mailed automatically to the students for their own records.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eInterviews\u003c/strong\u003e \u003cp\u003eThis study employed a qualitative research design, utilising semi-structured interviews to explore participants\u0026rsquo; experiences and perceptions of GenAI (Supplementary Information). Fifteen participants were recruited through email invitations. These participants represented a diverse cross-section of the MSc cohort. Each participant participated in a one-to-one, structured interview designed to elicit detailed responses related to their experiences with GenAI. Informed consent was obtained from all participants prior to the interview. Students were asked to confirm that they had read the participation sheet and consented to the interview. Automatic transcription built within teams was used to generate transcripts. To prepare the transcripts for analysis, an initial data cleaning process was undertaken. This involved the removal of redundant words, filler words (e.g., \u0026ldquo;um,\u0026rdquo; \u0026ldquo;like\u0026rdquo;), and other non-essential verbal tics, following best practices in qualitative data management (Gibbs, 2007). The cleaned transcripts kept the integrity of the participants\u0026rsquo; responses while ensuring clarity for later thematic analysis.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eThe data were analysed using thematic analysis, following the six-phase process outlined by Braun and Clarke (2019). Initially, the researchers familiarised themselves with the data by reading and re-reading the transcripts, followed by systematic coding of key phrases and patterns relevant to the research questions. These codes were grouped into broader themes, which were then reviewed and refined to ensure they accurately captured the participants\u0026rsquo; perspectives. The themes were clearly defined and named, and the final analysis was written up with supporting quotations from the participants. Throughout the process, reflexivity was maintained, and steps were taken to ensure the rigour and trustworthiness of the analysis by following guidelines from Lincoln and Guba (1985).\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e \u003cb\u003eEvaluation of initial attitudes to GenAI use.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo establish the student cohort's base-level understanding and attitudes to GenAI, students were polled during the initial GenAI seminar using a student response (clicker) system. The question followed taught content around how LLM\u0026rsquo;s and GenAI models operate. Students were asked, \u0026ldquo;How often do you use GenAI tools?\u0026rdquo; with 44% stating they use GenAI tools Always or Often (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). An open-text response was then used to capture which tools or models the students were using with ChatGPT was identified as the primary GenAI tool used by the student cohort.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eA set of Likert scale questions was employed to gauge students' initial perceptions and attitudes towards using GenAI in their studies. Students were asked to indicate whether they believed the following scenarios were acceptable uses of GenAI tools. The survey assessed participants' views on AI's usefulness across various academic tasks, such as writing assessments, understanding content, editing written work, and locating research articles (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe questionnaire reveals varying confidence levels in AI's capabilities across these areas.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eComprehension\u003c/strong\u003e \u003cp\u003emost respondents had a favourable view of AI's role in aiding content comprehension. 89% of participants agreed or strongly agreed that AI was helpful, with only 4% strongly disagreeing and 7% disagreeing. This suggests a strong confidence in AI's ability to assist in understanding complex topics.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eResearch articles\u003c/strong\u003e \u003cp\u003eAI was overwhelmingly viewed as beneficial for finding and understanding research articles. A substantial 86% of respondents agreed or strongly agreed with this capability, with only a small minority \u0026minus;\u0026thinsp;6% strongly disagreeing and 8% disagreeing - expressing doubts. This finding highlights the perceived value of AI tools in streamlining the research process and improving access to scholarly materials. Overall, the data suggest that while there is strong confidence in AI's ability to assist with content comprehension and research, students understood the ethical concerns around using GenAI in generating and editing written assessments.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eEditing written work\u003c/strong\u003e \u003cp\u003eOpinions on AI's ability to help with editing written content were more divided. While 43% of respondents agreed or strongly agreed that AI could assist in editing, 28% strongly disagreed, and 29% disagreed. This spread indicates that, although some users find AI useful for editing, a significant proportion remains unconvinced of its effectiveness.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eWrite assessments\u003c/strong\u003e \u003cp\u003eWhen asked if it was acceptable that AI could effectively write their assessments, a significant portion of respondents replied negatively, with 44% strongly disagreeing and 34% disagreeing. Only 22% of respondents agreed or strongly agreed that AI could perform assessment tasks. This highlighted that the majority of students understood concerns around academic integrity and highlighted considerable doubt about AI's ability to generate accurate and complete assessments independently.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eThese data informed the content and tone of taught sessions within the module.\u003c/p\u003e\n\u003ch3\u003eAssessment of perceived skills development\u003c/h3\u003e\n\u003cp\u003eA skills audit was conducted at the beginning and end of the module to assess how students perceived their skills around AI had progressed. Initially, the audit provides a baseline understanding of each student\u0026rsquo;s perceived existing knowledge and competencies. At the end of the modules, a second audit measures the self-reported improvements or changes in the skill set, thereby highlighting growth areas and identifying remaining gaps (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWithin the skills audit, students were asked to rate how important various research skills were to them. At the start of semester one, 77% of students rated knowing how GenAI operates as quite or extremely important, 79% thought prompt writing was quite or extremely important, 88% thought that understanding ethical use was quite or extremely important, and 90% reported that data integrity was quite or extremely important. No significant change in the value the students put on each aspect was recorded in the second iteration of the survey.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConfidence in How GenAI Operates\u003c/strong\u003e \u003cp\u003eAt the beginning of the semester, only 33% of students reported feeling \"Quite\" confident in their understanding of how GenAI operates, with 44% indicating \"Limited\" confidence and 23% reporting no confidence at all. By the end of the semester, confidence levels improved significantly, with 56% of students feeling \"Quite\" confident and only 19% reporting \"Not\" confident. This shift indicates a substantial increase in student understanding and confidence in operating GenAI tools.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConfidence in Writing Prompts\u003c/strong\u003e \u003cp\u003eInitially, 52% of students expressed \"Limited\" confidence in their ability to write prompts effectively, and only 21% felt \"Quite\" confident. By the end of the semester, the proportion of students feeling \"Quite\" confident more than doubled to 48%, while those with \"Limited\" confidence decreased to 33%. This trend suggests that instructional efforts in prompt writing were successful in boosting student confidence.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConfidence in the Ethical Use of AI\u003c/strong\u003e \u003cp\u003eEthical considerations are a critical aspect of AI usage. At the start of the semester, only 42% of students reported being \"Quite\" confident in their understanding of the ethical use of AI, with 18% indicating no confidence at all. By the semester's end, \"Quite\" confidence rose to 67%, and the proportion of students with no confidence dropped to just 10%. This improvement reflects the effectiveness of the curriculum in enhancing students' awareness and confidence in ethical AI practices.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConfidence in Data Protection Awareness\u003c/strong\u003e \u003cp\u003eData protection awareness showed similar trends. Initially, 45% of students felt \"Quite\" confident in their understanding of data protection, while 43% reported \"Limited\" confidence and 12% had no confidence. By the end of the semester, those \"Quite\" confident increased to 71%, and those \"Limited\" decreased to 23%, with only 6% reporting no confidence. This increase demonstrates significant progress in students' understanding and confidence in data protection principles.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eOverall, the data indicates a clear positive trend in student confidence across all assessed areas. These results suggest that the semester's instruction effectively enhanced students' understanding and confidence in key AI-related competencies.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStudents\u0026rsquo; voice\u003c/h2\u003e \u003cp\u003eDuring the third semester, fifteen one-on-one interviews were conducted to gauge the student cohort's perception of the changes made to the curriculum delivery in relation to their GenAI digital competence. Student researchers conducted all interviews to reduce data bias created by the perceived academic hierarchy between the MSc students and the interviewers. Each interview followed a specific question format as outlined in the supplementary data. The interview responses were transcribed and thematically coded, with three key themes being identified: 1) AI literacy and competence, 2) AI skill transferability, 3) Trust and Ethical Concerns.\u003c/p\u003e \u003cp\u003eThese themes are interdependent, forming a holistic experience of AI in academia. Increased utilisation leads to more learning and adaptation but also raises ethical concerns. AI's support is most effective when users have adapted to its use and trust it within ethical boundaries.\u003c/p\u003e \u003cp\u003eThe relationship between these themes is also cyclical. As students use AI and gain more experience, they refine their usage, adapt their learning, and continually reassess the ethical implications. This, in turn, influences how they continue to use AI and the support they look for from it.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eTheme 1 - AI literacy and competence\u003c/strong\u003e \u003cp\u003eStudents shared how the structured educational modules have significantly contributed to their understanding of GenAI tools. The experience has not only increased their confidence in using GenAI but has also improved their ability to use GenAI effectively in a range of tasks.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e \u003cem\u003e\"I've enhanced my skills in asking the right questions to AI and using tools like ChatGPT and others effectively after the [semester 1] module.\"\u003c/em\u003e \u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e\"My confidence in using AI has increased significantly thanks to the modules.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003eMany students use GenAI to assist with tasks such as summarising research papers, paraphrasing text, and tailoring CVs. AI is used to clarify complex concepts that students may not fully understand from lectures. Non-native English speakers use AI to help with language-related tasks, such as grammar checks and translation.\u003c/p\u003e \u003cp\u003e \u003cem\u003e\"It helps me with emails and checking my spelling because I am still learning English. It always translates for me or gives me synonyms, which makes it easier to write correctly.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e\"Sometimes the slang in articles is difficult to understand, so this tool is helpful for us, especially for those improving their English.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e\"I use AI to paraphrase and rephrase my writing, which helps me improve the quality of my work. It also provides feedback on grammar, which is crucial as I am still learning English.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eTheme 2 - AI skill transferability\u003c/strong\u003e \u003cp\u003eThe learning process includes understanding the limitations of GenAI, developing prompting skills, and recognising the broader potential of AI in their academic and professional lives.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e\"The modules have taught me how to use AI in a positive manner, whether for academic purposes or other tasks.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e\"The module taught me tricks like reading papers, creating outlines, and even preparing for job searches. Without AI, starting something like exam prep or coursework would have been more difficult.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003eStudents view AI as a valuable assistant that provides essential support in various areas. This includes language assistance for non-native speakers, simplifying complex tasks, guidance in learning new concepts, and feedback to improve their work. AI is seen as a tool that makes complex tasks more manageable by providing quick answers, summaries, and guidance. As students learn more about AI, they begin to use it more effectively as a supportive tool, refining their prompts and better understanding how to get the assistance they need.\u003c/p\u003e \u003cp\u003e \u003cem\u003e\"The module helped me understand the potential of AI beyond basic applications, such as writing essays. It showed me that AI could be a valuable tool if used correctly.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e\"The module taught us how to use AI tools effectively, not just for coursework but also for presentations and job searches. It\u0026rsquo;s like having a mentor guiding you through different tasks.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eTheme 3 - Trust and Ethical Concerns\u003c/strong\u003e \u003cp\u003estudents addressed concerns regarding GenAI's reliability and ethical use. Some students express doubts about the accuracy and reliability of AI, which affects their trust in using it for critical tasks, leading to a more informed and cautious approach to using these tools.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e\"I use it sometimes but I'm not a big fan, you know. ChatGPT, for example. I don't really trust it because sometimes it gives wrong answers.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003eStudents raised concerns about the ethical implications of using AI, especially regarding plagiarism and the originality of work. This negatively impacted their confidence in using GenAI tools for their studies.\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003e \u003cem\u003e\"Using AI to produce essays or coursework is unethical. We should use it only for understanding or improving our own work.\"\u003c/em\u003e \u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003e \u003cem\u003e\"We cannot put any personal info into AI because it could be discovered by others. I didn\u0026rsquo;t know it takes personal info and feeds it into the system.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003e\"I think I'm less confident in it after the [first semester] module because I didn\u0026rsquo;t realise things like plagiarism and data privacy.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003eStudents were also aware of the impact of AI on research integrity in the wider academic community, due to GenAI\u0026rsquo;s ability to generate \u0026ldquo;data\u0026rdquo; that hasn\u0026rsquo;t been empirically tested.\u003c/p\u003e \u003cp\u003e \u003cem\u003e\"Nowadays, some people use AI to generate data or images for research without doing the actual experiments, which could undermine scientific integrity.\"\u003c/em\u003e \u003c/p\u003e \u003cp\u003eTogether, these themes describe a complex and dynamic interaction between the students and GenAI. GenAI\u0026rsquo;s role in learning is not just about utility but is deeply intertwined with ethical considerations, the learning process, and the level of support it GenAI provides.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIntegrating GenAI into the MSc skills development program demonstrates both the potential and the challenges of embedding such technologies into higher education. Our findings suggest that while GenAI tools can significantly enhance personalised learning experiences and student engagement, their use also necessitates careful consideration of ethical implications, particularly regarding academic integrity and the development of critical cognitive skills.\u003c/p\u003e \u003cp\u003eStudents in this study and elsewhere ((Ngo, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) generally have a favourable opinion of using ChatGPT in education, citing benefits like time savings and personalised tutoring. However, they also identify barriers such as the inability to assess the quality and reliability of sources and the inability to cite sources accurately. One of the primary advantages observed was the role of GenAI in providing personalised feedback and support to students, particularly those from non-native English-speaking backgrounds. GenAI tools can enhance communicative practices for non-native English speakers and support students' language learning experiences both in the classroom and outside of it. These tools address issues such as lack of motivation, anxiety, limited authentic communication opportunities, and lack of personalised feedback. Similar observations about the support that GenAI can give learners are also reported in further case studies (Bedford et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Zadorozhnyy \u0026amp; Lai, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Here, the students in question used GenAI tools as part of a personalised English language enhancement course, reporting that AI-assisted learning was an effective way to identify issues with their writing. Strategies include utilising GenAI as a mentor for enhancing sentence structure, grammar, and spelling, providing personalised learning strategies and resources, recommending relevant language learning applications tailored to specific skills, and offering follow-up questions for comprehension checks (Bedford et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Zadorozhnyy \u0026amp; Lai, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). GenAI's ability to give students instant and personalised feedback on their writing and research tasks is, therefore, a promising development (Kasneci et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). However, concerns have been raised that prolonged use of GenAI tools may decrease higher-order thinking skills (Putra et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Richardson \u0026amp; Clesham, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Tenakwah et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTools like ChatGPT and ChatPDF were shown, here and elsewhere, to be instrumental in helping students navigate complex academic tasks (Elkhodr et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Kasneci et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), from summarising research papers to refining their writing. This aligns with the constructivist approach to learning, where students actively construct knowledge through interactions with these AI tools (Zhou \u0026amp; Schofield, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The ability of GenAI to offer tailored educational experiences is a significant advancement, making learning more accessible and engaging, especially for a diverse cohort (Guo et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Popenici \u0026amp; Kerr, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Such approaches allow students to receive tailored support while learning, enhancing engagement and educational outcomes aligning with individual learning trajectories (Bhutoria, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Hooda et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe increase in student confidence seen here across various GenAI-related competencies, such as prompt writing and ethical AI use, further underscores the value of integrating these technologies into the curriculum. The structured approach of embedding GenAI into core modules through the tip of the week, with \u0026ldquo;just-in-time\u0026rdquo; learning principles (Novak, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), allowed students to gradually build their skills, leading to a notable improvement in their self-assessed abilities by the end of the program. Just-in-time learning is rooted in delivering knowledge and skills precisely when needed (Welch, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). Embedding learning within relevant and meaningful situations, rather than teaching concepts in isolation, helps learners improve their understanding and application of concepts (Darling-Hammond \u0026amp; Snyder, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). This approach of applied learning is believed to enhance understanding, retention, and application of knowledge through real-life experiences and practical applications (Jach \u0026amp; Trolian, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eStudents generally view AI as a helpful tool that supports them in various tasks. This support is most beneficial when users are familiar with using AI effectively and trust that it can provide reliable assistance. The more users understand AI, the more they can depend on it for effective support. Learning to use AI involves asking the right questions and interpreting AI's responses. Students may initially rely heavily on AI for support. As they become more aware of ethical concerns, they might adjust how they use AI to ensure they uphold academic integrity. As students become more proficient and confident in using AI, they learn and adapt their strategies for incorporating these tools into their academic work. This theme emphasises the learning process associated with mastering AI tools and adjusting to their continually evolving capabilities.\u003c/p\u003e \u003cp\u003eDespite these benefits, the study also highlights significant concerns related to the ethical use of AI and the potential for academic dishonesty. The scepticism expressed by some students about the reliability of GenAI tools, coupled with the concerns about plagiarism and data privacy, indicates a need for ongoing education and clear guidelines on AI use in academic settings. The mixed responses regarding the appropriateness of AI in generating academic work suggest that while students appreciate the support AI can offer, they are also aware of the boundaries that must be maintained to ensure academic integrity. Clear guidelines for AI use in assessments can address these issues by setting out expectations and boundaries (Perkins, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This was achieved here through structured process-driven assessments allowing students to log GenAI usage (Smith \u0026amp; Francis, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) and setting aside seminar time to detail institutional guidelines. Moorhouse et al. (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), following their review of publicly available guidelines from the top 50 universities in the Times Higher Education World University Rankings 2023, noted a move to a redesign of assessments to effectively incorporate GenAI tools designing tasks that emphasise critical thinking and creativity (Bobula, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Chan, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Moorhouse et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Assessment methods include open-ended questions (Meir et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), project-based assessments, and working-world problem-solving tasks (Petrovska et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) that encourage original thought and application of knowledge, thereby reducing the potential for AI misuse (Meir et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). As GenAI technologies continue to evolve, there is a risk that students might bypass critical thinking processes, leading to a superficial understanding of the subject matter. The potential for over-reliance on AI tools, particularly in producing assessments, raises questions about the long-term impact on students' cognitive and creative skills (Richardson \u0026amp; Clesham, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Tenakwah et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). GenAI, use in learning, should complement rather than replace the essential human elements of learning.\u003c/p\u003e \u003cp\u003eThe success of this program can be attributed to the thoughtful integration of GenAI into the curriculum through experience mapping (Beard, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and the deliberate design of assessments (Smith \u0026amp; Francis, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). By aligning AI tools with core learning objectives and providing students with opportunities to critically engage with AI outputs, the program helped students develop a nuanced understanding of both the capabilities and limitations of GenAI.\u003c/p\u003e \u003cp\u003eHowever, the study also indicates that the effectiveness of such integrations depends heavily on the design and delivery of the curriculum. The cyclical relationship observed between students\u0026rsquo; increasing experience with GenAI and their evolving perceptions and ethical considerations highlights the need for continuous adaptation and refinement of teaching strategies. This iterative process is crucial in ensuring that GenAI tools enhance, rather than detract from, the educational experience.\u003c/p\u003e\n\u003ch3\u003eImplications for Future Research and Practice\u003c/h3\u003e\n\u003cp\u003eThis study provides valuable insights into the practical implementation of GenAI in higher education and opens several avenues for future research. The long-term effects of GenAI on student learning outcomes, particularly in terms of developing higher-order thinking skills, need to be explored. Additionally, further research is needed to evaluate the impact of AI on different student demographics, especially those who may be disproportionately affected by AI-related biases.\u003c/p\u003e \u003cp\u003eEducational institutions must prioritise developing comprehensive AI literacy programs for students and educators. These programs should foster a critical understanding of GenAI, encourage ethical use, and equip students with the skills necessary to navigate an increasingly AI-driven world. GenAI is being integrated into academic activities, serving as both a learning aid and a productivity tool. However, its use raises important questions about academic integrity and the potential for over-reliance. While AI tools are beneficial, students are rightly cautious about relying on them too heavily. This scepticism highlights the need for continued education on the ethical use of GenAI and the importance of critical thinking when interpreting AI-generated content. Structured learning is crucial in helping students gain confidence in using AI, which suggests that formal education on AI can significantly impact how effectively students utilise these tools.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eartificial intelligence\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGenAI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003egenerative artificial intelligence\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLLMs\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003elarge language models\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eGenAI declaration\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGTP4 (OpenAI) was used to create a broad paper outline and suggest content areas for discussion. The code used to generate the figures was drafted in GTP4 using dummy data before being edited and implemented in R, ensuring data integrity. Consensus was used to find and summarise relevant research articles for the introduction and discussion. Text editing was completed using Grammarly.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and/or analysed during the current study are not publicly available due to the confidential nature of the transcripts generated, but they are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was funded by a teaching and learning development grant from the College of Health, Wellbeing and Life Sciences (HWLS), Sheffield Hallam University, Sheffield, S1 1WB\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDPS conceived and designed the project, and DS, SM, CO, and CB acquired the data. All authors were involved in analysing and interpreting the data, and DPS, MML, and NJF wrote the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to acknowledge Dr Marjory Da Costa Abreu for her collaboration and input into the project. For the purpose of open access, the author has applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version of this paper arising from this submission.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAli, O., Murray, P. A., Momin, M., Dwivedi, Y. K., \u0026amp; Malik, T. (2024). The effects of artificial intelligence applications in educational settings: Challenges and strategies.\u003cem\u003e Technological Forecasting and Social Change, 199\u003c/em\u003e, 123076. \u003c/li\u003e\n\u003cli\u003eBeard, C. (2022). \u003cem\u003eExperiential learning design: Theoretical foundations and effective principles\u003c/em\u003e. Routledge. \u003c/li\u003e\n\u003cli\u003eBearman, M., \u0026amp; Luckin, R. (2020). Preparing university assessment for a world with AI: Tasks for human intelligence. In M. Bearman, P. Dawson, R. Ajjawi, J. Tai \u0026amp; D. Boud (Eds.), \u003cem\u003eRe-imagining university assessment in a digital world\u003c/em\u003e (pp. 49\u0026ndash;63). Springer International Publishing. https://doi.org/10.1007/978-3-030-41956-1_5\u003c/li\u003e\n\u003cli\u003eBedford, J., Kim, M., \u0026amp; Qin, J. C. (2024). Confidence enhancer, learning equalizer, and pedagogical ally. \u003cem\u003eUsing generative AI effectively in higher education: Sustainable and ethical practices for learning, teaching and assessment\u003c/em\u003e (1st ed., pp. 33). Taylor \u0026amp; Francis. https://doi.org/10.4324/9781003482918-17\u003c/li\u003e\n\u003cli\u003eBhutoria, A. (2022). Personalized education and artificial intelligence in the united states, china, and india: A systematic review using a human-in-the-loop model.\u003cem\u003e Computers and Education: Artificial Intelligence, 3\u003c/em\u003e, 100068. https://doi.org/10.1016/j.caeai.2022.100068\u003c/li\u003e\n\u003cli\u003eBobula, M. (2024). Generative artificial intelligence (AI) in higher education: A comprehensive review of challenges, opportunities, and implications.\u003cem\u003e Journal of Learning Development in Higher Education, \u003c/em\u003e(30)https://doi.org/10.47408/jldhe.vi30.1137\u003c/li\u003e\n\u003cli\u003eBrew, M., Taylor, S., Lam, R., Havemann, L., \u0026amp; Nerantzi, C. (2023). Towards developing AI literacy: Three student provocations on AI in higher education.\u003cem\u003e Asian Journal of Distance Education, 18\u003c/em\u003e(2), 1\u0026ndash;11. https://doi.org/10.5281/zenodo.8032387\u003c/li\u003e\n\u003cli\u003eChan, C. K. Y. (2023). A comprehensive AI policy education framework for university teaching and learning.\u003cem\u003e International Journal of Educational Technology in Higher Education, 20\u003c/em\u003e(1), 38. \u003c/li\u003e\n\u003cli\u003eChan, C. K. Y., \u0026amp; Colloton, T. (2024). \u003cem\u003eGenerative AI in higher education: The ChatGPT effect\u003c/em\u003e. Taylor \u0026amp; Francis. https://doi.org/10.4324/9781003459026\u003c/li\u003e\n\u003cli\u003eCrowther, M., Lim, W., \u0026amp; Crowther, M. A. (2010). Systematic review and meta-analysis methodology.\u003cem\u003e Blood, the Journal of the American Society of Hematology, 116\u003c/em\u003e(17), 3140\u0026ndash;3146. \u003c/li\u003e\n\u003cli\u003eDarling-Hammond, L., \u0026amp; Snyder, J. (2000). Authentic assessment of teaching in context.\u003cem\u003e Teaching and Teacher Education, 16\u003c/em\u003e(5-6), 523\u0026ndash;545. \u003c/li\u003e\n\u003cli\u003eElkhatat, A. M., Elsaid, K., \u0026amp; Almeer, S. (2023). Evaluating the efficacy of AI content detection tools in differentiating between human and AI-generated text.\u003cem\u003e International Journal for Educational Integrity, 19\u003c/em\u003e(1), 17. \u003c/li\u003e\n\u003cli\u003eElkhodr, M., Gide, E., Wu, R., \u0026amp; Darwish, O. (2023). ICT students\u0026rsquo; perceptions towards ChatGPT: An experimental reflective lab analysis.\u003cem\u003e STEM Education, 3\u003c/em\u003e(2), 70\u0026ndash;88. \u003c/li\u003e\n\u003cli\u003eFosnot, C. T. (2013). \u003cem\u003eConstructivism: Theory, perspectives, and practice\u003c/em\u003e. Teachers College Press. \u003c/li\u003e\n\u003cli\u003eGuo, H., Yi, W., \u0026amp; Liu, K. (2024). Enhancing constructivist learning: The role of generative AI in personalised learning experiences.\u003cem\u003e Proceedings of the 26th International Conference on Enterprise Information Systems (ICEIS 2024), 1\u003c/em\u003e, 767\u0026ndash;770. https://doi.org/10.5220/0012688700003690\u003c/li\u003e\n\u003cli\u003eHooda, M., Rana, C., Dahiya, O., Rizwan, A., \u0026amp; Hossain, M. S. (2022). Artificial intelligence for assessment and feedback to enhance student success in higher education.\u003cem\u003e Mathematical Problems in Engineering, 2022\u003c/em\u003e(1), 5215722. \u003c/li\u003e\n\u003cli\u003eHuxley-Binns, R., Lawrence, J., \u0026amp; Scott, G. (2023). Competence-based HE: Future proofing curricula. \u003cem\u003eIntegrative curricula: A multi-dimensional approach to pedagogy\u003c/em\u003e (pp. 131\u0026ndash;147). Emerald Publishing Limited. \u003c/li\u003e\n\u003cli\u003eJach, E., \u0026amp; Trolian, T. (2023). Supporting college student success through applied learning: Considering associations with average college grades, graduation in four years, and degree aspirations.\u003cem\u003e Journal of Postsecondary Student Success, 2\u003c/em\u003e(3), 75\u0026ndash;97. \u003c/li\u003e\n\u003cli\u003eJeltova, I., Birney, D., Fredine, N., Jarvin, L., Sternberg, R. J., \u0026amp; Grigorenko, E. L. (2007). Dynamic assessment as a process-oriented assessment in educational settings.\u003cem\u003e Advances in Speech Language Pathology, 9\u003c/em\u003e(4), 273\u0026ndash;285. \u003c/li\u003e\n\u003cli\u003eKasneci, E., Se\u0026szlig;ler, K., K\u0026uuml;chemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., G\u0026uuml;nnemann, S., \u0026amp; H\u0026uuml;llermeier, E. (2023). ChatGPT for good? on opportunities and challenges of large language models for education.\u003cem\u003e Learning and Individual Differences, 103\u003c/em\u003e, 102274. \u003c/li\u003e\n\u003cli\u003eLacey, M. M., \u0026amp; Smith, D. P. (2023). Teaching and assessment of the future today: Higher education and AI.\u003cem\u003e Microbiology Australia, 44\u003c/em\u003e(3), 124\u0026ndash;126. https://doi.org/10.1071/MA23036\u003c/li\u003e\n\u003cli\u003eLiang, W., Yuksekgonul, M., Mao, Y., Wu, E., \u0026amp; Zou, J. (2023). GPT detectors are biased against non-native english writers.\u003cem\u003e Patterns, 4\u003c/em\u003e(7), 100779. https://doi.org/10.1016/j.patter.2023.100779\u003c/li\u003e\n\u003cli\u003eLodge, J. M., de Barba, P., \u0026amp; Broadbent, J. (2023). Learning with generative artificial intelligence within a network of co-regulation.\u003cem\u003e Journal of University Teaching and Learning Practice, 20\u003c/em\u003e(7), 1\u0026ndash;10. \u003c/li\u003e\n\u003cli\u003eMeir, E., Pope, D., Abraham, J. K., Kim, K. J., Maruca, S., \u0026amp; Palacio, J. (2024). Designing activities to teach higher-order skills: How feedback and constraint affect learning of experimental design.\u003cem\u003e CBE\u0026mdash;Life Sciences Education, 23\u003c/em\u003e(1), ar1. \u003c/li\u003e\n\u003cli\u003eMoorhouse, B. L., Yeo, M. A., \u0026amp; Wan, Y. (2023). Generative AI tools and assessment: Guidelines of the world\u0026apos;s top-ranking universities.\u003cem\u003e Computers and Education Open, 5\u003c/em\u003e, 100151. \u003c/li\u003e\n\u003cli\u003eNgo, T. T. A. (2023). The perception by university students of the use of ChatGPT in education.\u003cem\u003e International Journal of Emerging Technologies in Learning (Online), 18\u003c/em\u003e(17), 4. \u003c/li\u003e\n\u003cli\u003eNoroozi, O., Soleimani, S., Farrokhnia, M., \u0026amp; Banihashem, S. K. (2024). Generative AI in education: Pedagogical, theoretical, and methodological perspectives.\u003cem\u003e International Journal of Technology in Education, 7\u003c/em\u003e(3), 373\u0026ndash;385. \u003c/li\u003e\n\u003cli\u003eNovak, G. M. (2011). Just‐in‐time teaching.\u003cem\u003e New Directions for Teaching and Learning, 2011\u003c/em\u003e(128), 63\u0026ndash;73. \u003c/li\u003e\n\u003cli\u003ePalmer, E., Lee, D., Arnold, M., Lekkas, D., Plastow, K., Ploeckl, F., Srivastav, A., \u0026amp; Strelan, P. (2023). Findings from a survey looking at attitudes towards AI and its use in teaching, learning and research.\u003cem\u003e ASCILITE Publications, \u003c/em\u003ehttps://doi.org/10.14742/apubs.2023.537\u003c/li\u003e\n\u003cli\u003ePerkins, M. (2023). Academic integrity considerations of AI large language models in the post-pandemic era: ChatGPT and beyond.\u003cem\u003e Journal of University Teaching \u0026amp; Learning Practice, 20\u003c/em\u003e(2)https://doi.org/10.53761/1.20.02.07\u003c/li\u003e\n\u003cli\u003ePetrovska, O., Clift, L., Moller, F., \u0026amp; Pearsall, R. (2024). Incorporating generative AI into software development education. Paper presented at the \u003cem\u003eProceedings of the 8th Conference on Computing Education Practice, \u003c/em\u003e37\u0026ndash;40. \u003c/li\u003e\n\u003cli\u003ePopenici, S. A., \u0026amp; Kerr, S. (2017). Exploring the impact of artificial intelligence on teaching and learning in higher education.\u003cem\u003e Research and Practice in Technology Enhanced Learning, 12\u003c/em\u003e(1), 22. \u003c/li\u003e\n\u003cli\u003ePutra, F. W., Rangka, I. B., Aminah, S., \u0026amp; Aditama, M. H. (2023). ChatGPT in the higher education environment: Perspectives from the theory of high order thinking skills.\u003cem\u003e Journal of Public Health, 45\u003c/em\u003e(4), e840\u0026ndash;e841. \u003c/li\u003e\n\u003cli\u003eRichardson, M., \u0026amp; Clesham, R. (2021). Rise of the machines? the evolving role of artificial intelligence (AI) technologies in high stakes assessment.\u003cem\u003e London Review of Education, 19\u003c/em\u003e(1), 1\u0026ndash;13. \u003c/li\u003e\n\u003cli\u003eRodriguez, J. G., Hunter, K. H., Scharlott, L. J., \u0026amp; Becker, N. M. (2020). A review of research on process oriented guided inquiry learning: Implications for research and practice.\u003cem\u003e Journal of Chemical Education, 97\u003c/em\u003e(10), 3506\u0026ndash;3520. \u003c/li\u003e\n\u003cli\u003eRomova, Z., \u0026amp; Andrew, M. (2011). Teaching and assessing academic writing via the portfolio: Benefits for learners of english as an additional language.\u003cem\u003e Assessing Writing, 16\u003c/em\u003e(2), 111\u0026ndash;122. \u003c/li\u003e\n\u003cli\u003eRudolph, J., Tan, S., \u0026amp; Tan, S. (2023). ChatGPT: Bullshit spewer or the end of traditional assessments in higher education?\u003cem\u003e Journal of Applied Learning and Teaching, 6\u003c/em\u003e(1), 342\u0026ndash;363. \u003c/li\u003e\n\u003cli\u003eSallam, M. (2023). ChatGPT utility in healthcare education, research, and practice: Systematic review on the promising perspectives and valid concerns.\u003cem\u003e Healthcare, 11\u003c/em\u003e(6), 887. https://doi.org/10.3390/healthcare11060887\u003c/li\u003e\n\u003cli\u003eSchreiber, N., They\u0026szlig;en, H., \u0026amp; Schecker, H. (2016). Process-oriented and product-oriented assessment of experimental skills in physics: A comparison. Paper presented at the \u003cem\u003eInsights from Research in Science Teaching and Learning: Selected Papers from the ESERA 2013 Conference, \u003c/em\u003e29\u0026ndash;43. \u003c/li\u003e\n\u003cli\u003eSmith, D., \u0026amp; Francis, N. (2024). Process not product in the written assessment. \u003cem\u003eUsing generative AI effectively in higher education\u003c/em\u003e (1st ed., pp. 115\u0026ndash;126). Routledge. https://doi.org/10.4324/9781003482918-17\u003c/li\u003e\n\u003cli\u003eSok, S., \u0026amp; Heng, K. (2024). Opportunities, challenges, and strategies for using ChatGPT in higher education: A literature review.\u003cem\u003e Journal of Digital Educational Technology, 4\u003c/em\u003e(1), ep2401. https://doi.org/10.30935/jdet/14027\u003c/li\u003e\n\u003cli\u003eSotiriadou, P., Logan, D., Daly, A., \u0026amp; Guest, R. (2020). The role of authentic assessment to preserve academic integrity and promote skill development and employability.\u003cem\u003e Studies in Higher Education, 45\u003c/em\u003e(11), 2132\u0026ndash;2148. https://doi.org/10.1080/03075079.2019.1582015\u003c/li\u003e\n\u003cli\u003eTenakwah, E. S., Boadu, G., Tenakwah, E. J., Parzakonis, M., Brady, M., Kansiime, P., Said, S., Ayilu, R., Radavoi, C., \u0026amp; Berman, A. (2023). Generative AI and higher education assessments: A competency-based analysis.\u003cem\u003e Res.Sq, \u003c/em\u003ehttps://doi.org/10.21203/rs.3.rs-2968456/v2\u003c/li\u003e\n\u003cli\u003eWelch, R. (2010). How just in time learning should become the norm! Paper presented at the \u003cem\u003e2010 Annual Conference \u0026amp; Exposition, \u003c/em\u003e15.649. 1\u0026ndash;15.649. 12. \u003c/li\u003e\n\u003cli\u003eZadorozhnyy, A., \u0026amp; Lai, W. Y. W. (2023). ChatGPT and L2 written communication: A game-changer or just another tool?\u003cem\u003e Languages, 9\u003c/em\u003e(1), 5. \u003c/li\u003e\n\u003cli\u003eZhou, X., \u0026amp; Schofield, L. (2024). Using social learning theories to explore the role of generative artificial intelligence (AI) in collaborative learning.\u003cem\u003e Journal of Learning Development in Higher Education, \u003c/em\u003e(30)https://doi.org/10.47408/jldhe.vi30.1031\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Sheffield Hallam University","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":"Generative Artificial Intelligence, Higher Education, Process base Assessment, Assessment, AI","lastPublishedDoi":"10.21203/rs.3.rs-5204546/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5204546/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe arrival of Generative Artificial Intelligence (GenAI) into higher education has brought about significant transformations in assessment practices and pedagogical approaches. Large Language Models (LLMs) powered by GenAI present unprecedented opportunities for personalised learning journeys. However, the emergence of GenAI in higher education raises concerns regarding academic integrity and the development of essential cognitive and creative skills among students. Critics worry about the potential decline in academic standards and the perpetuation of biases inherent in the training sets used for LLMs. Addressing these concerns requires clear frameworks and continual evaluation and updating of assessment practices to leverage GenAI's capabilities while preserving academic integrity. Here, we evaluated the integration of GenAI into a year-long MSc program to enhance student understanding and confidence in using GenAI. Approaching GenAI as a digital competency, its use was integrated into core skills modules across two semesters, focusing on ethical considerations, prompt engineering, and tool usage. The assessment tasks were redesigned to incorporate GenAI, which takes a process-based assessment approach. Students' perceptions were evaluated alongside skills audits, and they reported increased confidence in using GenAI. Thematic analysis of one-to-one interviews revealed a cyclical relationship between students' usage of GenAI, experience, ethical considerations, and learning adaptation.\u003c/p\u003e","manuscriptTitle":"Embedding Generative AI as a digital capability into a year-long MSc skills program","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-10-08 10:31:02","doi":"10.21203/rs.3.rs-5204546/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":"ff9e3d82-cc75-45fe-854a-b2bee70d8ac3","owner":[],"postedDate":"October 8th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-10-08T10:31:02+00:00","versionOfRecord":[],"versionCreatedAt":"2024-10-08 10:31:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5204546","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5204546","identity":"rs-5204546","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}