Is using AI for university or school work cheating?

It depends entirely on your institution, and that is not a dodge — policies differ between universities, between departments, and sometimes between two assignments on one module. Some permit AI for planning or revision but not drafting, some require you to declare what you used, and some forbid it in assessed work. Read your own policy and the assignment brief, then ask the person marking the work whenever the wording is unclear.

How do I use AI to revise without letting it do the thinking?

Give it the questioning role and keep the recall for yourself. Close the book and have it test you on the chapter. Write your own explanation first, then ask what is missing or wrong. Ask it to argue against your thesis so you can answer the objection. A rough check: if the session ends and you still could not reproduce the material unaided, the tool did the part that was yours.

Can I trust an AI explanation of something on my syllabus?

Not without checking it. These models produce fluent text whether or not the claim underneath is correct, and a wrong date or a garbled formula is worse during revision than no note at all, because you will remember it. Use the explanation to get unstuck, then verify it against your lecture material or set text. Pointing a tool at your own documents narrows the problem without removing it.

Is it acceptable to use AI to improve my English if it is my second language?

Language support and content authorship are treated as different things in most places, though not every policy draws the line identically. Correcting grammar in sentences you wrote yourself is usually viewed differently from having a model produce the argument for you. Because that varies, check your institution's rules and the brief rather than assuming. Grammar and translation tools also flatten voice, so read the result before you submit it.

Topic hub

AI for Students

Students hear two contradictory things about AI: that it is the end of real learning, and that refusing to touch it will leave them behind. Neither is much help at eleven at night with a deadline. This page gathers everything on the site that is genuinely useful for studying, and says plainly where the line sits.

One idea sorts almost every use of AI in study into helpful or harmful, and it has nothing to do with which tool you pick. AI helps you when it creates difficulty, and it hurts you when it removes difficulty. Everything below is arranged around that distinction.

Learning happens while your brain is doing work it cannot yet do comfortably: pulling a fact back from memory without the page in front of you, explaining a mechanism in your own words, noticing the exact moment you run out of explanation. That last feeling — the one where you thought you understood something and discover you can only recite it — is not a sign the session is going badly. It is the session working. Anything that reliably removes it leaves you with the sensation of understanding, which is far easier to manufacture than the real thing and feels almost identical until an exam asks a question in an unfamiliar order.

Using AI to create difficulty looks like this. You ask it to quiz you on a chapter and close the chapter. You write your own explanation of a concept, paste it in, and ask what a marker would say is missing or wrong. You ask for the same idea explained at three levels and notice which level you can reproduce without looking. You ask it to argue the strongest case against your essay thesis so you can answer that objection before someone else raises it. You turn your own messy lecture notes into flashcards and then answer them from memory. In every one of those, you are still the one doing the retrieval; the tool is only holding the questions and being available at an hour when nobody else is.

Using AI to remove difficulty looks like this. You have it summarise the reading you were assigned so you can speak in the seminar without doing it. You have it draft the essay and then adjust the wording until it sounds like you. You ask for the answer to a problem instead of the next hint. This feels productive, because output appears and the deadline is met, but the work you handed over is the work that was the point of setting it. The cost is deferred rather than avoided, and it usually arrives during a closed-book exam, a viva, or the first follow-up question in an interview.

There is a second, more mundane problem: these systems produce fluent, confident text that is sometimes simply wrong. A study aid that invents a date, garbles a formula, or fabricates a citation is worse than no study aid at all, because you will revise from it and remember it. Treat anything a model tells you about your subject as a claim to check against your actual course material — the lecture slides, the set text, the marker in front of you — rather than as a source in its own right. Feeding your own sources in, as document tools let you do, reduces this problem but does not remove it.

On academic integrity, one plain paragraph. Rules about AI vary enormously — between institutions, between departments inside the same institution, and sometimes between two assignments on the same module. Some allow AI for planning and revision but not for drafting. Some require a written declaration of what you used and how. Some prohibit it entirely for assessed work. No website can tell you which of those applies to you, and this one will not try. Find your own institution's policy, read the wording in each assignment brief, and ask the person marking the work when it is ambiguous — before you submit, not after. Nothing on this site is guidance on getting around those rules, and none of it is legal or academic advice.

What follows is the material on the site that fits the principle above: a free course on how these systems actually work, guides written with students in mind, prompt templates designed to test you rather than to write for you, and independent write-ups of the tools people reach for most — including what each of them is bad at.

Start with the course

Knowing why these tools fail is what stops you trusting them in the wrong places. A beginner-friendly walkthrough of what today's AI systems are, how they learn from data, and the vocabulary you need to follow any AI conversation with confidence.

AI Fundamentals: How Modern AI Actually Works

4 lessons · about 10 min of reading · free, no sign-up

Read the guides

The first two are written for study specifically. The last two are about checking what a model tells you, which matters more when you are revising from it than when you are drafting an email.

Best Free AI Tools for Students

Which AI tools genuinely help with studying, what their free tiers actually include, and the academic-integrity line you shouldn't cross.

Using AI to Study Without Cheating Yourself

AI can genuinely accelerate learning — or quietly replace the thinking that makes learning stick. The difference is in how you use it.

Free AI Tools: What You Can Do Without Paying

Most major AI tools have a real free tier. Here's what's genuinely possible at zero cost — and where the limits usually bite.

How to Fact-Check AI Output

A repeatable process for checking AI-generated claims — including which parts to check first when you don't have time to check everything.

AI Hallucinations: Why AI Makes Things Up

AI systems can state false information with total confidence. Here's why that happens and how to protect yourself from it.

Prompts that make you do the work

Each template explains why its wording produces what it does, what good output looks like, and the mistakes that lead to weak results. The study ones are deliberately built so that you supply the recall and the tool supplies the questioning.

The full prompt library covers writing, planning, and job-hunting tasks as well.

Tools students actually reach for

Independent descriptions based on documented behaviour and ordinary use, not formal testing or benchmarks. Each write-up lists real limitations alongside the strengths, and free tiers and features change often enough that the official site is the only reliable place to check what you would get today.

NotebookLMStudying or researching from a specific set of documents — asking questions, generating summaries, and getting cited answers grounded in your own sources.PerplexityResearch questions where you want a synthesized answer with sources to verify.ChatGPTDrafting emails and documents, brainstorming, summarizing long text, and answering general questions.ClaudeLong-document analysis, detailed writing, and coding assistance that needs a large amount of context at once.GrammarlyCatching grammar and clarity issues in emails, documents, and messages as you write them.DeepLTranslating text or documents between languages while keeping natural, readable phrasing.

None of these is the right answer on its own, and the choice depends on the task: grounding answers in your own documents, or searching with sources you can open, or general drafting and questioning, or tidying language. Two tools used well beat six tabs.

Key terms

Three words worth knowing before you rely on any of this for revision — each defined in plain English, with an honest note on what the term does not cover.

Questions students ask

Is using AI for university or school work cheating?: It depends entirely on your institution, and that is not a dodge — policies differ between universities, between departments, and sometimes between two assignments on one module. Some permit AI for planning or revision but not drafting, some require you to declare what you used, and some forbid it in assessed work. Read your own policy and the assignment brief, then ask the person marking the work whenever the wording is unclear.
How do I use AI to revise without letting it do the thinking?: Give it the questioning role and keep the recall for yourself. Close the book and have it test you on the chapter. Write your own explanation first, then ask what is missing or wrong. Ask it to argue against your thesis so you can answer the objection. A rough check: if the session ends and you still could not reproduce the material unaided, the tool did the part that was yours.
Can I trust an AI explanation of something on my syllabus?: Not without checking it. These models produce fluent text whether or not the claim underneath is correct, and a wrong date or a garbled formula is worse during revision than no note at all, because you will remember it. Use the explanation to get unstuck, then verify it against your lecture material or set text. Pointing a tool at your own documents narrows the problem without removing it.
Is it acceptable to use AI to improve my English if it is my second language?: Language support and content authorship are treated as different things in most places, though not every policy draws the line identically. Correcting grammar in sentences you wrote yourself is usually viewed differently from having a model produce the argument for you. Because that varies, check your institution's rules and the brief rather than assuming. Grammar and translation tools also flatten voice, so read the result before you submit it.

Where to go next

If your questions keep coming back badly phrased rather than badly answered, the Prompt Engineering hub explains why specificity changes results, and Prompt Engineering for Real Work turns that into a habit. When study turns into applications, using AI for a job search covers what these tools can and cannot do for a CV, and the tools directory lists everything written up here with its limitations attached.