AI Tools for Students: ChatGPT, Claude, Gemini, Elicit and More

Start with the rules: what is allowed?

Policies on generative AI differ widely between universities, departments and even individual assignments. Common models are: AI use prohibited, AI allowed for specific tasks (for example brainstorming or language polishing) with disclosure, or AI fully allowed with disclosure. Some assessments, such as exams or reflective essays, are usually AI-free.

Before you use any tool, check: the assignment brief, the module handbook, your university’s academic integrity policy and any guidance from your library. If anything is unclear, ask your instructor in writing. “I did not know” rarely helps in an academic misconduct case. Our article on AI in academic writing explains the principles in more depth.

Overview: AI tools and what students use them for

This overview describes typical uses as of late 2026. Features, models and terms change quickly; check the provider’s current information and your university’s licence (many universities provide an institutional version of a chatbot with better data protection).

Useful and legitimate uses (where allowed)

Used well, AI tools act like a patient tutor rather than a ghostwriter. Examples that many universities accept, sometimes with disclosure:

• Understanding: “Explain the difference between reliability and validity with an example from education research.”
• Practising: generating quiz questions from your own lecture notes, then answering without help.
• Planning: brainstorming possible research questions, then refining them yourself with your supervisor.
• Feedback: asking for comments on the logic of your outline or the clarity of a paragraph you wrote.
• Language: checking grammar and spelling, especially if English is not your first language (check whether this needs disclosure).
• Searching: using Elicit, Consensus or Perplexity as a starting point, then reading the actual papers via your library.

Limits: hallucinations and fabricated references

Large language models generate text that sounds right. They do not look up facts in a reliable database unless connected to search, and even then they can misread sources. Typical problems:

• Invented references: realistic-looking author names, titles, journals and DOIs that do not exist. Always find the source in your library catalogue or via the DOI before citing it.
• Wrong attributions: a real paper cited for a claim it does not make.
• Outdated or oversimplified information, especially on statistics, law and medicine.
• Bias: one-sided perspectives that reflect the training data.
• Data protection: do not paste interview transcripts, personal data or unpublished research into tools that are not approved by your university.

How to disclose and cite AI use

If your university allows AI use, it will usually require some form of transparency. Common forms are an AI use statement (what tool, for what, how you checked the output), a note in the methods section of empirical work, and in some cases a citation of the tool.

APA style (general guidance from the APA Style team): treat the tool’s developer as the author and cite the tool with its version, in the pattern: Developer. (Year). Tool name (Version) [Large language model]. URL. For ChatGPT, the author element is the company that develops it. In the text, describe how you used it and, where relevant, quote the prompt and the relevant part of the response; longer exchanges can go in an appendix.

MLA style (general guidance from the MLA): do not treat the AI tool as an author; describe the prompt in the title element and name the tool, version, company, date and URL.

Universities may have their own rules that override these style guides, for example a mandatory declaration form. Follow those first. Professor Digital’s AI log records every AI-assisted step automatically and creates the declaration in the format your university requires.

A responsible workflow in five steps

A simple routine keeps you on the safe side:

• 1. Check the policy for the specific assignment.
• 2. Decide the purpose before you open a tool: understand, practise, plan or get feedback, not “write my essay”.
• 3. Verify every factual claim and every source in the original.
• 4. Write yourself: use AI feedback to revise your own text instead of pasting generated text.
• 5. Document prompts, outputs and how you used them, then disclose as required.

Professor Digital and AI

Professor Digital is built on the principle “help, not ghostwriting”: it provides building blocks such as research question suggestions, outlines, questionnaire items and feedback, but not ready-to-submit chapters. Every AI-assisted step is recorded in your AI log, and you can upload your university’s AI guidelines so that advice is aligned with them. Our free tools for citations, statistics and planning work without AI.