Unit 1 of 5

Ethical Use for Learners

Before the tricks, the ground rules: what counts as help, what counts as cheating, and how to stay honest.

about 12 minutes

By the end of this unit you can

Tell the difference between AI as a tutor and AI as a ghostwriter
Spot hallucinations and verify claims before you submit
Protect your own and other people's private data
Declare and cite AI use the way your institution expects

Start here

AI is a study partner, not a substitute

Every tool in this workshop can make you faster. None of them can make you understand something for you. The line that matters is simple: AI may help you learn, but the thinking you are assessed on must be yours. If a tool produced the reasoning, the writing or the analysis that your grade rewards, you have crossed the line.

Most universities and schools now allow AI for learning support and restrict it for assessed output. Your first job in any subject is to read that subject's policy, because it differs between courses, and "my other lecturer allowed it" has never worked as a defence.

The replacement test

Before using AI on a task, ask: if my lecturer watched me do this, would I be comfortable explaining it? If the honest answer is no, don't do it.

Decide fast

The green, amber and red zones

Use this grid as a quick check whenever you are unsure. Amber is the biggest zone in real student life. It is not forbidden, it just needs permission and disclosure.

Green: clearly fine

Asking AI to explain a concept in simpler words
Generating practice questions and quizzing yourself
Asking for feedback on structure or clarity of your own draft
Summarising your own lecture notes to revise faster
Turning a topic into a study plan or timeline

Amber: ask first, always declare

Using AI to rewrite or polish sentences you wrote
Translating your own writing into another language
Letting AI generate an essay outline you then write from
Using AI-generated code or data analysis in an assignment
Using AI to find sources you have not read yet

Red: academic misconduct

Submitting AI text as if you wrote it
Using AI in a closed assessment or exam that forbids it
Pasting confidential or copyrighted material into a public chatbot
Citing sources AI invented without checking they exist
Having AI complete a group member's or peer's assessed work

Accuracy

Hallucinations: confident, fluent and wrong

Chatbots predict plausible text. Plausible is not the same as true. The most common student disaster is a citation that does not exist: a real-sounding author, a real journal, a fake article. Markers check these, and a fabricated reference is treated as fabrication of evidence.

1
Ground the model in real sources
Prefer NotebookLM or an uploaded PDF over open chat when facts matter. Grounded tools quote your own material.
2
Ask for the source, then open it
A link you did not click is not a source. Verify the title, author and year in your library database or Google Scholar.
3
Cross-check with a second tool
If Gemini and ChatGPT disagree on a fact, treat both as unverified until a human-authored source settles it.
4
Never cite the chatbot as evidence
Cite the original study. AI is a tool you disclose, not an authority you quote for facts.

Force the model to separate fact from guess

Answer my question below. Then add two sections:
1. "High confidence": claims you can support with well-known, verifiable sources, each with author, year and title.
2. "Low confidence / may be wrong": anything you are inferring or unsure about.
Do not invent citations. If you cannot verify a source exists, say so explicitly.

My question: [paste your question]

Use this on any factual answer you plan to reuse in coursework.

Privacy

What never goes into a chatbot

Free consumer AI tools may use your conversations to improve their models. Treat every chat window as semi-public.

Your student ID, matric number, address, phone number or IC/passport details
Other people's names, grades, health details or interview transcripts
Unpublished research data, industry placement material or anything under NDA
Full exam papers or copyrighted textbooks your licence does not allow you to upload

Anonymise before you paste

Replace names with "Participant A", remove identifiers, and use your institution's licensed accounts (often Gemini for Education or ChatGPT Edu) where training on your data is switched off.

Disclosure

How to declare AI use properly

Declaring AI use protects you. A short, specific statement is almost always accepted; a vague or missing one is what triggers suspicion.

Example declaration

"I used ChatGPT (GPT-5, Oct 2026) to generate practice questions on cell respiration and to get feedback on the clarity of my introduction. All arguments, sources and final wording are my own. I used NotebookLM to summarise the three lecture PDFs listed in my references."

Draft your AI-use statement

I used AI in the following ways for this assignment: [list what you actually did, e.g. brainstormed ideas, generated practice questions, checked grammar].
Write a 2-3 sentence academic integrity declaration in formal English that states exactly which tool and version I used, what I used it for, and confirms the analysis and final wording are mine. Do not overstate or understate my use.

Keep your receipts

Save your chat links, drafts and version history. If your work is ever questioned, a document history showing your writing evolve is the strongest evidence you have.

Check yourself

Five questions before you submit

Can I explain every sentence in this work without looking at it?
Have I opened and read every source I cite?
Does my course policy allow the way I used AI here?
Have I declared my AI use, if declaration is required?
Did I put anything private or confidential into the tool?

If all five answers feel comfortable, you are ready for the rest of this workshop, where the goal is to use these tools so well that you actually learn more, not less.