On this site
I use AI tools the way I use a dictionary and a search engine: as an assistant during research and revision. Specifically, they help me find sources to read, suggest alternative phrasings, and check drafts for unclear passages.
They do not write the resources. Every page here is drafted, argued, and revised by me, and every factual claim is checked against a source I have read. Where a page carries my name it means I stand behind what it says, which is a claim no generated text can make.
This matters for a practical reason. Language models produce confident, fluent, plausible writing about citation rules and disciplinary conventions, and they get details wrong in ways that are invisible unless you already know the answer. A student following that advice loses marks. So the checking is not a formality.
In your own work
Your institution's policy governs, and policies vary widely and change often. Read yours before using any tool, and if it is ambiguous, ask your tutor rather than guessing. Nothing here overrides it.
Beyond compliance, some general observations from coaching students through this.
- AI is genuinely useful for things you can verify: explaining a concept you will then check, generating counter-arguments to test your position, reformatting a reference list you will then proofread.
- It is unreliable for citations. Fabricated references with plausible authors, titles, and page numbers are common, and markers check.
- It is a poor substitute for thinking. The struggle to articulate an argument is where the learning is, and outsourcing it produces a paper you cannot defend in a viva or build on next semester.
- Whatever you use, you remain responsible for every claim you submit under your name.
Disclosure
If that ever changes, this page changes with it, and the change is dated. A policy that only ever describes best practice is not a policy.