Usually not — most UK universities allow using AI to brainstorm ideas or plan a structure, as long as you write the essay yourself in your own words. It generally becomes misconduct when AI generates the actual text you submit.
The exact rule is set by your course, so check your module's assessment policy. That wording, not a general principle, is what a misconduct panel would apply.
Why so many students feel guilty about this
In 2026 AI is close to universal in UK higher education — HEPI reports that 94% of UK undergraduates use generative AI for assessed work in some form. When something is that widespread and the rules are that unevenly explained, the result is a lot of people who aren't cheating and still feel like they might be.
The guilt usually isn't about wrongdoing. It's about not knowing where the line sits.
A test that actually works
Could you explain and defend every argument in your essay, out loud, without the AI in front of you? If yes, it's your work. If the ideas and the words are the model's and you tidied them, that's the danger zone.
This test maps onto how the rules are actually enforced. Assessment measures your understanding; a plan you critiqued, rejected half of and rebuilt demonstrates understanding. Paragraphs you accepted wholesale do not — which is also why in-person vivas and oral defences are making a comeback in 2026.
Where departments usually draw it
- Usually fine: asking for an explanation of a concept, generating a list of angles to consider, sanity-checking a structure, getting your grammar and spelling corrected — then writing the essay yourself.
- Usually not fine: generating paragraphs and submitting them as your own words, having AI construct your argument, or paraphrasing AI output to disguise where it came from.
- Genuinely varies: using AI to summarise a source you then cite, and using it to draft a section you rewrite heavily. Some courses require you to declare this; some prohibit it. Check.
One more thing worth knowing: a plan is not a source, and an AI-generated claim is not evidence. If the model hands you a fact, find and cite the real thing — fabricated references are a much faster route to a misconduct case than the AI use itself.
When in doubt, ask
A two-line email to your module leader — “Is it acceptable to use AI to help plan the structure of this assignment, provided I write it myself?” — settles this far faster than worrying about it, and gives you something in writing if it ever matters.
Sage is SafeGrade's academic advisor — it knows your university and course, and can talk you through where the line sits for your assignment before you submit.
Try SafeGrade free →No card. No catch. First scan free.