University AI misconduct with unclear rules

Short answer

Yes, in principle you can still be penalised, but each of the three situations you mention — genuinely unclear rules, rules that changed after you did the work, and a tutor telling you it was allowed — can be a strong ground of defence or appeal. Whether a penalty is fair depends heavily on the exact wording of the rules that applied to your specific assessment at the time you did it, and on what you were actually told. The key legal and procedural principles are fairness, clarity of the rules, and the burden being on the university to prove misconduct on the balance of probabilities.

How universities treat unauthorised AI use

Most universities in England and Wales now treat unauthorised or undeclared use of generative AI as a form of academic misconduct, usually classed as false authorship or unfair means, on the basis that AI-generated content is not your own work. Penalties can range from a mark reduction or a capped resit to failing the module or, in serious or repeated cases, more severe outcomes.

Crucially, this is not general law but the university’s own regulations and your student contract. The university must follow its own published academic misconduct procedures, and those procedures must be applied fairly.

Where “unclear rules” helps you

To find misconduct, a university normally has to show, on the balance of probabilities, that AI use was not permitted for that assessment and that you knew or should reasonably have known that. If the rules were genuinely ambiguous, that undermines the case against you.

However, be aware of a common trap. Many universities have a default rule that if the position on AI is unclear or silent, students must assume AI is not permitted unless expressly allowed. Birmingham City University and others state this explicitly. So “the rules were unclear” is only a strong defence if there was no such default position, or if the guidance actively pointed in different directions, or if the ambiguity was in the assessment brief itself rather than something you failed to check.

The strongest version of this argument is where the assessment brief, module handbook and general policy contradicted each other, or where a reasonable student reading the actual instructions could genuinely have believed AI was allowed.

Where “the rules changed later” helps you

This is often the strongest point. A penalty should be judged against the rules that were in force at the time you produced and submitted the work, not against a policy introduced afterwards. Applying a new, stricter AI rule retrospectively to work done before it existed is generally unfair and is a legitimate ground of appeal or complaint.

You will need to pin down the dates: when you did and submitted the work, and when the relevant AI policy or assessment guidance was published or amended. If the tightened rule post-dates your submission, argue that it cannot fairly be used to penalise you.

Where “a tutor said it was allowed” helps you

If a member of staff with apparent authority told you, in writing or verbally, that AI use was permitted for that assessment, and you relied on that in good faith, it is unfair to penalise you for following that instruction. This goes to the heart of whether you had any dishonest or improper intent and whether the university’s own communications were consistent.

The practical difficulty is proof. Written evidence — an email, a message on the virtual learning environment, a post in a module forum, or lecture slides — is far stronger than an unrecorded conversation. If it was verbal, note who said it, when, in what session, and whether anyone else heard it. Also consider whether the tutor was speaking about that specific assessment or making a general comment, because the university may argue the permission did not extend to your particular task.

The role of the Office of the Independent Adjudicator

The OIA (the independent higher education complaints ombudsman for England and Wales) has issued specific casework guidance on AI and academic misconduct. It expects providers to have clear rules, to make findings on the balance of probabilities, to give clear reasons for both the finding and the choice of penalty, and to explain why lesser penalties were not appropriate. If a university’s AI rules were unclear, applied inconsistently, or changed without proper notice, that is exactly the kind of concern the OIA scrutinises. You can normally take a complaint to the OIA only after you have exhausted the university’s internal procedures and obtained a Completion of Procedures letter.

Practical next steps

1. Gather the evidence now. Save the exact assessment brief, module handbook, and any AI or academic integrity policy as they appeared when you did the work, ideally with dates or version numbers. Screenshots and downloaded PDFs are useful because online pages can be edited.

2. Preserve any communication from staff about AI, especially anything saying it was allowed. Retrieve emails, VLE announcements, forum posts and lecture materials before access is removed.

3. Reconstruct a timeline. Note when you did the work, when you submitted, and when any AI policy was introduced or changed.

4. Engage properly with the university’s academic misconduct process. Ask for the specific rule you are alleged to have breached and the evidence relied on. Put your points in writing: that the applicable rules were unclear or silent, or were changed afterwards, or that staff told you AI was permitted, and that you acted in good faith.

5. Use the internal appeal if the decision goes against you. Appeals usually require a ground such as procedural unfairness, a material irregularity, or new evidence, all of which map onto the arguments above. Check and meet the appeal deadline, which is often short.

6. If the internal process is exhausted and you remain dissatisfied, obtain the Completion of Procedures letter and consider a complaint to the OIA within its time limit.

7. Check whether your students’ union offers a free advice or advocacy service. They frequently represent students in misconduct hearings and know how the local process operates.

Key missing facts that change the answer

The strength of your position turns on several things I do not yet know: the precise wording of the assessment brief and AI policy that applied to your specific task at the time; whether there was a default “assume not permitted unless stated” rule; the exact dates of any policy change relative to your submission; and whether any staff permission was in writing and clearly covered that assessment. If you can tell me what the brief and policy actually said, when you submitted, and what the tutor said and how, I can give you a much more targeted view of how likely a penalty is to stand and what your best line of challenge is.