Ethical Use of AI Tools When Writing a Thesis or Dissertation in Africa

What the rules actually say about AI and authorship

Postgraduate students across Africa are using AI tools for literature searches, drafting help, and data organisation faster than most graduate schools have written formal policy to match. Where a standalone university policy does not exist yet, the clearest reference point is the Committee on Publication Ethics (COPE) position statement on authorship and AI tools, which most journals and many graduate schools now follow. COPE states plainly that AI tools cannot be listed as an author on any research output. Authorship carries responsibility and accountability that only a person can hold. An AI tool cannot take responsibility for the accuracy of a thesis chapter, cannot hold copyright, and cannot be held accountable if something in the work turns out to be wrong. That single rule shapes almost everything else about how you can use AI while writing a thesis or dissertation.

Where AI tools help, safely

• Summarising and organising literature you have already found and read, to help you spot patterns across dozens of papers faster.

• Improving grammar, clarity, and structure in a draft you wrote yourself.

• Generating a first-pass outline or structure for a chapter you then write and rewrite in your own words.

• Formatting references once you have verified each one is real and accurate.

• Explaining a statistical method or concept to you as a study aid, separate from anything you submit.

Where supervisors draw the line

• Generating your findings, analysis, or original argument. That is the core intellectual contribution examiners assess, and it has to come from you.

• Fabricated or unverifiable citations. AI tools sometimes produce plausible-looking references that do not exist. Check every source an AI tool gives you against the original before it goes in your bibliography.

• Submitting AI-generated text as your own without disclosure, even after you edit it. Most universities treat this as an academic integrity violation similar to plagiarism.

• Using AI on data you are not permitted to share externally, including participant data covered by your ethics approval. Many AI tools process input on external servers, which can breach research ethics or data protection commitments you made to participants.

Why disclosure matters more than the tool itself

Universities across Africa are still catching up on AI-specific graduate policy, and standalone graduate-school AI rules remain rare outside a small number of institutions. Where a formal policy does not exist yet, your existing plagiarism and research ethics policy still applies, and disclosure is the safest default. Treat AI assistance the way you would treat help from a research assistant or a proofreader: name it, describe what it did, and keep the intellectual work and final judgment yours. COPE's position has been adopted by many publishers, so a disclosure habit you build now will match what journals expect if you go on to publish from your thesis.

Before you submit

• Confirm your specific graduate school's or journal's AI policy in writing, since rules differ by department and by publisher.

• Keep a record of your prompts and AI outputs for any chapter where you used a tool, in case your supervisor or examiner asks.

• Re-verify every fact, statistic, and citation an AI tool touched, against the original source.

• Disclose AI use in your methods, acknowledgements, or a dedicated declaration, following your institution's template if one exists.