What role may GenAI play in academic misconduct or poor academic practice? 

This guidance has been created to support Unit Convenors/markers who find suspected academic misconduct or poor academic practice and for all involved in the formal investigation of suspected academic misconduct. It should be read in conjunction with Regulations for Students 19. 

Published on: 04/06/2026 · Last updated on: 04/06/2026

Introduction

It is important to remember that unethical and/or unauthorised use of GenAI is not in itself a type of academic misconduct, rather it is a means by which students may engage in one or more of the types of academic misconduct that the University does recognise.

Unethical and/or unauthorised use of Gen AI can also play a role where students require further guidance around academic integrity because they have demonstrated poor academic practice. Poor academic practice is not academic misconduct.

Definitions of different levels of misconduct

Poor academic practice does not meet the level of academic misconduct. It is defined as conduct that demonstrates a lack of understanding and shows limited evidence of plagiarism, self-plagiarism or collusion. It is important to note that all other types of misconduct cannot be classified as poor academic practice.

In which areas of offence or poor practice is GenAI most likely to be used?

GenAI is most likely to be involved in cases that involve:

• Fabrication – this includes fabrication of references, data, evidence or experimental results and otherwise false, misleading or negligent representation of information which is included in work submitted for assessment.
• Plagiarism – presenting someone else’s work or ideas as the student’s own. This includes paraphrasing without acknowledgement, whether by the student themselves or by a GenAI tool.
• Collusion – working with someone else, or a GenAI tool, on an assessment which is intended to be the student’s sole work. This includes sharing work and/or answers with other persons.

We explore these issues in the case studies at the bottom of this page.

What’s the threshold between an academic misconduct offence or a case of poor academic practice?

Ultimately, this will come down to academic judgment and the evidence you are able to collate.

Poor academic practice may be judged to have occurred when a student’s work shows a lack of understanding leading to limited evidence of plagiarism or self-plagiarism, or limited evidence of collusion during a group work activity. All other types of misconduct must be classified as either moderate or severe misconduct, as specified in Regulations for Students 19.

You must use your academic judgment to determine whether the evidence you have collected indicates academic misconduct or poor academic practice. You must consider factors such as:

• the extent of the misconduct
• the amount of evidence available

and whether the student has:

• intended to deceive,
• gained an unfair advantage,
• engaged in academic misconduct before.

Investigating suspected academic misconduct involving inappropriate GenAI use may feel complex. However, the case studies below should help you understand that the investigation process is similar to when GenAI is not considered as part of the case. In cases of fabrication, and in some cases of plagiarism there is likely to be clear evidence. It is less likely but possible in the case of collusion involving GenAI misuse.

Note: academic misconduct does not need to be proved beyond reasonable doubt rather the standard of proof is upon the balance of probabilities, in other words, something is more likely to have happened than not. It is also important that procedures are followed accurately as outlined in our Regulations for Students 19.

Suggestions as to which evidence to collect can be found here, along with guidance on the proper use of similarity checking software and the interpretation of its results.  

If a student is deemed to have demonstrated poor academic practice, what actions might follow?

The case should be returned  to the Unit Convenor, who might:

• stipulate the student retakes the academic integrity test
• advise them to make an appointment with the Skills Centre – https://www.bath.ac.uk/campaigns/skills-for-success-programme/
• Provide the student with feedback to help them avoid unethical use of GenAI in future, if this has been involved.

How do I uphold academic integrity on a distance learning course?

Upholding academic integrity in distance learning may be more challenging, but it is by no means impossible. By their very nature, many online programmes are assessed predominantly or entirely by coursework and there may be few or no built-in opportunities for the kind of synchronous interactions that provide a baseline sense of a student’s ability. Then too, there may be a large pool of markers who are therefore unable to recognise any unexplained changes in individual writing style and markers and students will have no informal in-person interactions or shared time in a lab that can help staff build familiarity with students’ work.

The pool of available evidence to draw from when forming a suspicion or investigating misconduct may be smaller, however, it is not absent. Previous assignments remain a valuable resource to draw on for comparison purposes where plagiarism or collusion is suspected, and where fabrication of references has occurred, the evidence is as stark as on a campus-based course.

At the University of Bath any suspected academic misconduct offence must be investigated following the processes described in our Regulations for Students 19. A Presentation of Learning, if the student is willing to participate, is likely to be a key part of your investigation and our guidance lays out how this may be undertaken remotely.

It is important to recognise that these challenges in upholding academic integrity are not unique to distance learning students but apply, to varying degrees, to all large cohort units. They highlight the need for a diversified assessment portfolio where the limitations of one assessment type are balanced by the strengths of another. For further detail on this, please consult our article about minimising academic misconduct through assessment design.

Illustrative case studies

The case studies described below are presented for educational and illustrative purposes only and do not depict real events or people. Often, there will be multiple suspected offences in a single assessment but there is no threshold of number of suspected offences required for an investigation to commence.

Case Study 1: Fabrication of Experimental Data and References

A postgraduate student, Alex, submits a dissertation in a biomedical sciences programme. The project involves lab-based experiments measuring the efficacy of a new compound on cell regeneration. The dissertation includes detailed experimental results and cites several peer-reviewed articles to support the methodology and findings. One of the examiners notices the experimental results are unusually consistent and lack expected variability. This might in itself warrant an investigation. The examiner also notices that there are two articles cited that they are unfamiliar with, and which do not appear in any academic database. They report this to the unit convenor.

1. Preliminary Review

The Unit Convenor reviews the work: They note the overly consistent experimental results and the unverified sources.

2. Formal Inquiry

The Director of Studies agrees with the Unit Convenor’s assessment and starts an investigation.

The Director of Studies reviews:

• a) the evidence from the Unit Convenor.
• b) any record of previous academic misconduct by the student.

The Director of Studies consults the course team: The academic supervisor is asked whether they have reviewed drafts of the student’s work at intervals.

The Director of Studies checks for any laboratory records: These may indicate how much lab work was undertaken by Alex.

Student Notification:

Alex is informed of the allegations and is provided with any relevant evidence. Alex is invited to attend a meeting with the Director of Studies, where the meeting focusses on:

• details of the suspected academic misconduct
• the evidence
• the processes for investigation
• possible penalties

Alex does not admit the offence and so this is followed by a Presentation of Learning, which gives the student the opportunity to demonstrate that the submitted work is their own.

3. Evidence base

• Lab Records: No lab usage is recorded for key dates.
• Reference Verification: Two references appear to be fictional and three are misquoted or misrepresented.
• Raw Data Request: Alex was asked to provide raw data files. The files are either missing or inconsistent with the results reported in the dissertation.

4. Outcome

• The Director of Studies conclusion: There is evidence of severe misconduct. The Director of Studies believes that Alex has engaged in fabrication of experimental data and references.  The case is referred to the Head of Department to review.
• The Head of Department reviews the evidence: agrees that the evidence warrants a Board of Inquiry. The Head of Department presents their findings, and those of the Director of Studies, to the Chair of the Board of Studies who convenes a Board of Inquiry into Academic Misconduct.
• Board of Inquiry findings:
• Fabrication Confirmed: The panel concludes that Alex fabricated both experimental data and references.
• Intentional Misconduct: Given the scale and nature of the falsification, the panel determines it was deliberate and constitutes severe academic misconduct.

6. Penalties:

• Academic Penalty: Alex is withdrawn from their course (penalty 10). Alex remains eligible for exit awards under the relevant assessment regulations.
• Record Annotation: A note is added to their academic record indicating misconduct.
• Appeal Rights: Alex is informed of the right to review the decision of the Board of inquiry on limited grounds, as described within Regulations for Students 19.

Case Study 2: Plagiarism of AI-Generated Content

Jordan, an undergraduate student on a Politics programme, submits, in semester 1, a final-year essay on ‘The Role of Myth in Modern Political Narratives.’ The essay is polished and well-structured; however, a tutor notices the inclusion of what appear to be prompts to a GenAI tool as well as speech directly addressing the reader. Jordan has not acknowledged the use of GenAI in the creation of the essay.

This is brought to the unit convenor’s attention, who suspects the use of generative AI tools (e.g., ChatGPT, Gemini) without proper attribution. The Unit Convenor is concerned that Jordan has plagiarised material.

1. Preliminary Review

• Comparison with Past Work: The unit convenor compares the essay with Jordan’s recent assignments. They do not contain direct speech or phrases that read like GenAI prompts.

2. Formal Inquiry

• Director of Studies decision: The Director of Studies agrees with the Unit Convenor’s assessment and believes there is sufficient evidence to warrant an investigation.
• Student Notification: Jordan is informed of the suspicion and invited to a meeting with the Director of Studies as part of their investigation, during which Jordan acknowledges using an AI tool to draft the essay. He accepts that by not referencing this use he has engaged in plagiarism.

3. Evidence base

• Unusual phrasing: The essay contains apparent GenAI prompts and direct speech, whilst neither are referenced.
• Student admission: Jordan has admitted use of GenAI.

4. Outcome

• The Director of Studies conclusion: Jordan submitted AI-generated content without appropriate attribution, contravening the academic integrity statement he signed when he submitted the work. The case is referred to the Head of Department to confirm that moderate academic misconduct has occurred, as the case meets the criteria that firstly, the student has not previously been found to have engaged in academic misconduct and secondly, if the error had gone undetected, this would have resulted in a substantial unfair advantage to them since the student would have been credited with high quality work that had actually been produced by GenAI.

5. Penalties

• Educational Remedy: The Head of Department confirms that on the balance of probabilities and with the evidence available, Jordan has engaged in moderate academic misconduct and applies penalty 4. Jordan must resubmit the work. The mark is capped at the minimum pass mark.
• Record Annotation: A note is added to their academic record indicating academic misconduct.
• Appeal Rights: Jordan is informed of the right to review the decision of the Head of Department on limited grounds as described within Regulations for Students 19.

Case Study 3: Collusion with GenAI During an Online Exam

Taylor, a second-year student in a computer science programme, sits a timed, in-person, invigilated online exam on algorithms and data structures. The exam is open-book, with a specific book on this field allowed. The exam explicitly prohibits the use of generative AI tools. The exam is controlled via a lockdown browser and all phones and other internet-enabled devices must be left in bags at the side of the room. During the exam, the invigilator notices a candidate wearing glasses and making small rhythmic movements whilst apparently staring into space. Rather than challenge the candidate then and there the invigilator makes a note of the candidate’s identity for follow up later.

The invigilator suspects Taylor may have been using smart glasses during the exam and reports their suspicions to the unit convenor.

1. Preliminary Review

Exam Script Analysis: The unit convenor reads the exam script and finds that several answers contain terminology and phrasing not used in lectures or course materials. The unit convenor also compares Taylor’s answers with their previous assessment submissions and finds a considerable shift in style and vocabulary.

2. Formal inquiry

• Director of Studies decision: The Director of Studies agrees with the Unit Convenor’s assessment and believes there is sufficient evidence to warrant an investigation.
• Technical Logs: The student’s browser logs at the time of the exam are reviewed.
• Policy Check: The exam information provided to students ahead of the assessment clearly indicates that GenAI use is prohibited and define it as a breach of exam regulations, which is a form of academic misconduct.
• Student Notification: Taylor is informed of the suspicion and invited to meet with the Director of Studies and to a subsequent Presentation of Learning.

3. Evidence Base

• Technical Logs: Browser logs show that no prohibited websites were accessed.
• Student Interview: Taylor admits to accessing ChatGPT during the exam, claiming they misunderstood the rules.

4. Outcome

• The Director of Studies concludes that Taylor used GenAI assistance during the exam, thereby committing collusion and breaching examination regulations. They classify it as severe academic misconduct and present their findings to the Head of Department who agrees and requests that the Chair of the Board of Studies convenes a Board of Inquiry.

5. Penalties

• Academic Penalty: Taylor receives a zero for the exam (penalty 5), consequently fails the unit and is required to retake it. Whilst the University does not consider a student’s intentions or any mitigating factors when determining whether they have engaged in academic misconduct, they do consider any mitigating factors when determining penalties for academic misconduct.  
• Educational Remedy: Taylor must complete an academic integrity test before the end of this current stage. It will be recommended that they resit the academic integrity test before resitting their exam.
• Record Annotation: A formal note is added to the student’s academic record.

Case Study 4: Suspicion of AI Plagiarism Without Conclusive Evidence

Morgan, a first-year student in a Social Policy programme, submits an essay titled ‘Universal Basic Income and the Future of Welfare Reform.’ The essay demonstrates a level of policy analysis and terminology that would not normally be seen in year 1 (level 4) of this course. Coupled with this there are some abrupt changes in tense and style within the work. This raises the suspicion that chunks of text may have been pasted in and not properly attributed.

1. Preliminary Review

• Comparison with Past Work: The unit convenor reviews Morgan’s previous essays and notes a significant shift in style and complexity.
• Reference Check: All sources are verifiable and cited correctly.

2. Formal Inquiry

• The unit convenor consults the Director of Studies: who decides to start an investigation.
• Student Meeting: Morgan is invited to meet the Director of Studies, and he explains what will happen in the investigation.  The Director of Studies realises that a Presentation of Learning is required and invites Morgan to discuss the essay at a subsequent meeting.
• Presentation of Learning: Morgan claims to have worked independently, using lecture notes, policy reports, and academic databases, but does admit limited use of GenAI. She is insistent that she read all the references AI offered her. She is apologetic and said this was not properly cited as she is unfamiliar with referencing conventions.

3. Outcome

• The Director of Studies conclusion: Morgan has shown poor academic practice. The Director of Studies refers the matter back to the Unit Convenor as per Section 2, of Regulations for Students 19 .
• Learning Conversation: Instead of pursuing disciplinary action, the unit convenor writes to Morgan using the provided letter template. They then engage Morgan in a reflective conversation about academic integrity and the ethical use of GenAI.
• Skill Development: Morgan is offered support to develop academic writing skills and is encouraged to attend workshops on responsible AI use.
• Educational remedy: Morgan must also retake the academic integrity training and test before the end of her current stage. The Unit Convenor asks that she takes it as soon as possible in order to help avoid this issue again.