There is no single rule. What is allowed depends on the course, the assignment
and sometimes the individual instructor. Here is how to find out where the line is.
Last reviewed 12 November 2026
A great many AI misconduct cases involve students who did not think they were cheating. One
instructor encourages chatbots for brainstorming; the next treats a grammar tool as
unauthorised help. The same action can be fine on Monday's assignment and a violation on
Friday's.
Where the rules actually are
Check these in order. The more specific source normally wins.
The assignment brief. Any statement about tools on the task itself.
The course syllabus or handbook. Look for a section on AI, academic
integrity or permitted resources.
Department or programme rules.
The institution's academic integrity policy. If it says nothing specific
about AI, older rules on "unauthorised assistance" or "work that is not your own" usually
apply.
If none of these answers your question, ask the instructor by email before you start, and
keep the reply.
A rough spectrum
These are generalisations. Your course may draw the line anywhere along this list.
Usually accepted
Spell-check and basic grammar correction built into a word processor.
Citation managers and reference formatters.
Searching for sources, as long as you read and verify them yourself.
Depends on the course
Asking a chatbot to explain a concept you are studying.
Brainstorming topics or getting feedback on a draft you wrote.
Grammar tools that rewrite whole sentences for tone or clarity.
Translating your own writing from another language.
Generating an outline that you then write up yourself.
Usually treated as misconduct unless expressly allowed
Submitting generated text as your own, edited or not.
Having a tool paraphrase a source, or someone else's work, to pass it off as yours.
Using AI in an exam or test where it has not been permitted.
Using AI to produce data, quotations or references that you have not checked.
Invented references are a separate problem. Chatbots can produce
citations that look real and do not exist. Submitting them can be treated as fabrication, which
many policies regard as more serious than unauthorised help.
Three questions that settle most cases
Was it allowed for this task? If you cannot point to something that says
yes, assume you need to ask.
Did you say you used it? Undisclosed use is what most policies punish.
Many courses that allow AI require you to state how you used it.
Is the thinking yours? If the assignment exists to assess your argument,
your analysis or your writing, and a tool supplied that part, it is likely to be a
problem.
Protecting yourself
Ask early and keep the answer in writing.
If AI use is allowed, add a short note saying which tool you used and for what.