The Ethical Use of AI in Academic Research

by admin Uncategorized

By Micah Asamba

First Published on keynearesearch.com

You are working on a master’s thesis in Nairobi. Your university library closes at 6 PM. The journals you need sit behind paywalls that cost $40 per article. Your supervisor is overwhelmed, responses take weeks, and the methodology you’re trying to understand was written for a reader with resources you simply do not have.

This is not a niche problem. It is the everyday reality for thousands of graduate students across Kenya and the broader African continent.

AI tools — used honestly — can change this. The question is not whether to use them. The question is how to use them without compromising the integrity that makes your research worth anything in the first place.

What AI Can Legitimately Do for You

Understand Dense Literature

You can paste a complex methods section into Claude or ChatGPT and ask it to explain how a particular statistical model works. That is not cheating. That is using a tool to understand something, exactly the way you would ask a classmate or read a textbook. The understanding you build is yours.

Use it to decode jargon, trace an argument’s logic, or get plain-language explanations of concepts your supervisor assumes you already know.

Search and Synthesize

AI cannot replace a systematic literature review — but it can help you map a field quickly. Ask it to explain the major debates in your area, the key theorists, the methodological fault lines. Then verify everything. AI hallucinates citations. Treat every reference it gives you as unconfirmed until you have found the actual paper yourself in Google Scholar, Semantic Scholar, or through your institution’s library access.

A practical rule: AI tells you what to look for. You find the real thing.

Write Better, Not Instead

AI is excellent at helping non-native English speakers write more clearly. If English is your third language and you have a clear idea but a clunky sentence, asking an AI to improve your phrasing is legitimate. What is not legitimate is asking it to generate your argument for you.

The distinction matters: your ideas, your analysis, your contribution — AI helps you express them more clearly.

Think Through Your Methodology

Stuck on whether to use a case study or a comparative design? AI is a patient sounding board. Describe your research problem and ask it to walk you through the tradeoffs. It will not replace your supervisor’s judgment — but it can help you arrive at your next supervision meeting with sharper questions.

Where the Ethical Lines Are

Ghost-writing is plagiarism

If you submit text that AI wrote and present it as your own, you are committing academic fraud. This applies whether or not your institution’s plagiarism checker catches it. The problem is not detection. The problem is that you are claiming intellectual work you did not do.

More practically: your examiners will assess whether you can defend your work in a viva. If an AI wrote your literature review, you will not be able to.

Citations you have not read are dishonest

AI will confidently cite papers that do not exist. Even when the paper does exist, the AI’s summary of it is often subtly wrong. Citing a source you have not read — especially one you sourced from AI — is a form of deception. It also produces bad research.

Read the papers. Or if you genuinely cannot access them, say so in your methodology and work with what you have.

Disclose what you used

Many universities in Kenya are still developing AI policies. Check yours. Where policies are unclear, err toward transparency. A short note in your methodology — “AI tools were used to assist with literature mapping and language editing” — is professional and honest. Hiding it is not.

The field is moving fast. Honesty now protects your reputation later.

The Specific Opportunity for Kenyan Researchers

Let’s be direct about something. Researchers in high-income countries have access to things you often do not: supervisors who respond within 48 hours, library budgets that cover most journals, peers working on adjacent problems, writing centers, statistics help desks.

AI partially compensates for that gap. Not perfectly. But meaningfully.

A PhD student in Eldoret can now access the same explanations of multilevel modeling that a student at LSE gets from a methods seminar. A master’s student in Mombasa can get detailed feedback on the logical structure of her argument at 11 PM when no one else is available. That is real.

Use it for that. Use it to level a field that is not level.

What you cannot do is use it as a substitute for your own thinking. African scholarship has its own perspectives, its own questions, its own contributions to make to global knowledge. Those can only come from you. AI trained mostly on Western academic literature cannot produce insights grounded in Kenyan contexts, Kenyan data, or Kenyan lived experience. That is your comparative advantage. Protect it.

A Practical Framework

Before you use an AI tool for any research task, ask yourself three questions:

1. Am I using this to understand or to avoid understanding? Using AI to decode a paper: fine. Using AI to skip reading the paper: not fine.

2. Would I be comfortable telling my supervisor exactly what I did? If the answer is no, you probably should not do it.

3. Does this output require me to verify it before I use it? Almost always: yes. Do the verification.

Tools Worth Knowing

A few that are free or low-cost and useful for researchers:

  • Semantic Scholar — free academic search with AI-assisted summaries
  • Consensus.app — AI that synthesizes evidence across papers on a question
  • Elicit — AI research assistant for literature review
  • Unpaywall — browser extension that finds legal free versions of paywalled papers
  • Research4Life — if you are at a Kenyan university, check whether your institution is registered; it provides access to thousands of journals at no cost
  • Claude, ChatGPT, Gemini — for explanation, thinking, and writing assistance; not for citation or generation of research content

The Bottom Line

AI is not going to write your thesis for you. Not ethically, and not well. But used honestly, it can make you a more effective researcher — one who understands more, reads more efficiently, writes more clearly, and thinks through problems more rigorously.

That is what it is for.

The researchers who will benefit most are not the ones who outsource their thinking to AI. They are the ones who use it to sharpen their thinking, then do the hard work themselves.

You have real research to do. Questions that matter, in contexts that deserve serious scholarship. Use every legitimate tool available to do that work well.

keynearesearch.com supports graduate researchers across Kenya and East Africa. For more on research methods, academic writing, and navigating graduate study, explore our other resources.