Learning with AI: What's Allowed and What's Not

AI tools are now part of student life whether universities like it or not. The question is not whether to use them — it is how to use them in a way that actually helps you learn and does not put your degree at risk.

1. AI as a learning partner (the right way)

Used well, an AI assistant is a patient tutor available at 3 a.m. that never gets tired of your questions. Genuinely useful applications:

• Ask for explanations of concepts you did not understand in the lecture. Follow up until it clicks: "Can you explain that differently?" "Give me a concrete example."
• Socratic practice: ask the AI to quiz you on a topic. Try to answer from memory; use its feedback to find gaps.
• Generate practice questions on a topic before an exam, then answer them on paper without the AI's help.
• Summarise your own notes — paste in what you wrote, ask the AI to restructure or highlight what is missing. You still did the intellectual work.
• Get feedback on your draft: share your essay outline or a paragraph and ask "What is weak in this argument?" You then rewrite it yourself.
• Check your understanding: explain a concept to the AI in your own words and ask it to point out what is wrong or incomplete.

In all of these cases you are the one thinking; the AI is the mirror and the sparring partner.

2. What is generally allowed vs. not allowed

Rules vary by institution and assignment — always read the specific instructions. As a general guide:

Generally allowed	Generally not allowed
Using AI to understand source material	Submitting AI-generated text as your own writing
Generating practice questions to self-test	Asking AI to answer exam questions for you
Getting feedback on your own draft	Having AI rewrite your draft and submitting that
Summarising notes you wrote yourself	Using AI where the assignment explicitly forbids it
Checking grammar/spelling on your own text	Fabricating or leaving in AI-invented citations
Brainstorming ideas and then developing them yourself	Hiding AI use when disclosure is required

The core principle is simple: if the work you hand in does not represent your own thinking, it is misconduct — regardless of what tool produced it.

3. The grey zone and why university policy always wins

Many situations are genuinely ambiguous: using AI to translate your notes, having it restructure a paragraph you wrote, or generating a bibliography template. Different lecturers and institutions draw the line in different places.

Always check your university's specific policy on AI use. Many now publish explicit guidelines, and some require you to disclose AI use in a declaration at the end of your work. When in doubt, ask your lecturer before submitting — not after.

If a policy says "no AI tools," that includes the helpful uses above. Treat the restriction as a deliberate pedagogical choice: the skill-building happens precisely when you have to struggle through without a shortcut.

Tip: If you used AI in any part of your research or writing process, note it — even briefly. Many institutions now welcome transparent disclosure and will penalise concealment far more harshly than the use itself.

4. AI hallucinates — verify everything

AI language models produce fluent, confident text even when they are wrong. Common failure modes:

• Fabricated citations: an AI may invent plausible-looking author names, journal titles, and page numbers that do not exist. Every reference must be verified against a real source before you cite it.
• Outdated information: models have a training cutoff and may not know about recent legal changes, scientific findings, or policy updates.
• Subtle factual errors: a statistic, a date, a definition — stated confidently and incorrectly. Cross-check anything factual against a textbook, a database, or a primary source.

The rule is simple: if it goes in your work, you are responsible for it. "The AI told me" is not a defence.

5. Practical habits

A few habits that keep AI use productive and safe:

1. Always start work before opening an AI chat. Even ten minutes of your own thinking first gives you a baseline to compare against.
2. Keep AI conversations separate from your notes. Paste only your own words into your document.
3. When you ask for an explanation, test yourself immediately afterwards without looking at the chat.
4. Save your AI conversations if your university may ask you to show your working.
5. Treat a surprisingly polished AI answer as a prompt to understand it, not a text to copy.

---

Related guides on student-online cover smart exam preparation and time management for students.