What you will learn
- Choose study tasks where AI assistance can support learning rather than replace it
- Check AI-generated facts, quotations, calculations and references against reliable sources
- Protect personal information and follow course rules on acknowledgement and permitted use
Start with the learning goal
Before opening an AI tool, decide what you need to learn. A useful goal might be to understand a difficult concept, compare two explanations, generate practice questions or receive feedback on a draft you have written. A vague request such as “do my assignment” does not identify the skill you are meant to practise.
AI works best as a learning aid when you remain active: ask questions, inspect the reasoning, compare the answer with course material and explain the result in your own words.
Useful study tasks and their limits
- Explain a concept: ask for a beginner explanation, then compare it with your textbook or lecture notes.
- Practise recall: request short quiz questions, answer them without help and use feedback to find gaps.
- Explore a topic: ask for search terms, sub-questions or relevant subject areas, then locate original sources yourself.
- Review your writing: ask for comments on clarity or organization, while keeping the ideas and final decisions under your control.
- Understand code: request an explanation of a small example and then test your own understanding by changing it or writing a new solution.
These are possible learning uses, not automatic permission to use AI in every assessment. A teacher or institution may restrict some uses because of the learning outcomes being assessed.
Check your course and institution's rules first
Rules differ between courses, assignments and institutions. One assignment may allow idea generation but not AI-written paragraphs; another may ask students to use an AI tool and critique its output. Read the assessment brief and ask the instructor when the rule is unclear.
Where disclosure is required, state which tool you used, what you used it for and how you checked or revised its contribution. Do not present generated material as independent work when the assignment requires your own writing or reasoning. Keep drafts, notes and source records when your institution asks for them.
Integrity rule
Your submission should reflect your own understanding and comply with the exact instructions for that task. General guidance cannot replace your course's policy.
Verify facts instead of trusting fluent wording
A generated answer may sound certain while containing an incorrect date, invented detail, outdated rule or missing qualification. This can happen because a language model generates a plausible response; fluent wording is not proof that a claim has been checked.
Use a simple verification routine:
- Underline factual claims that affect your conclusion.
- Find an authoritative source, such as the original study, official dataset, technical standard or government guidance.
- Open the source and confirm that it actually supports the claim.
- Check the date, scope, method and limitations. A result for one country, age group or year may not apply everywhere.
- Rewrite your notes in your own words and keep the source details for citation.
References and quotations need extra checking
AI tools can produce references that look realistic but contain the wrong title, author, publication date, page or link. Never include a citation just because a tool supplied it. Search for the document in the publisher's catalogue, journal website, official repository or library database and inspect the original.
For quotations, verify the exact words and page or section in the source. If you cannot locate the original passage, do not present it as a quotation. For numerical claims, check the dataset, units, sample and calculation rather than copying a number from a generated summary.
Recognize bias, missing context and over-reliance
AI output can reflect gaps or imbalances in available information. It may leave out local context, present one interpretation as universal or summarize a debate too simply. Ask what evidence is missing, whose perspective is not represented and whether a different source changes the conclusion.
Do not let the tool make a decision that you cannot explain. If you cannot describe why an answer is correct without looking at the generated response, return to the lesson material and work through the concept again.
Protect personal and confidential information
Do not paste passwords, identity numbers, private student records, medical details, unpublished research, confidential employer documents or another person's personal information into a public AI service. Check the tool's privacy settings and your institution's approved options before sharing data. When possible, replace sensitive details with a fictional example.
Also check the terms that apply to copyrighted material, research data and intellectual property. Permission to use a tool does not automatically mean you can upload every document to it.
Worked example: checking an AI-generated study claim
A student asks an AI assistant to explain whether a particular programming language is always faster than another. The response gives a confident yes and lists two benchmark figures. Instead of copying the statement, the student should ask what workload was tested, which versions were compared, what hardware was used and whether the figures came from an actual benchmark.
The student then looks for the benchmark report or documentation, checks its method and date, and compares another credible source. If the evidence shows that performance depends on the task and implementation, the student should write that qualified conclusion rather than repeat the tool's broad claim.
Practice task
- Choose a topic from a current class and write one specific question you want to understand.
- Ask an AI tool for a plain-language explanation or five practice questions, if your course permits it.
- Select two factual claims and verify each against an original or authoritative source.
- Record whether the generated response missed a limitation or gave an unsupported reference.
- Write a short explanation from your own understanding and acknowledge AI use if your course requires it.
Quick review
- AI can support explanation, practice and feedback, but does not replace learning or judgment.
- Assignment rules determine what AI assistance is permitted.
- Verify facts, quotations, calculations and references in the original sources.
- Protect personal, confidential and unpublished information.
- Keep your own notes and be transparent where acknowledgement is required.
Further reading
- UNESCO: Guidance for generative AI in education and research.
- NIST: Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile.
- Higher Education Commission, Pakistan: Framework on use of Generative AI Tools in Higher Education Institutions.
- University of Dundee: Core policy on Generative AI in education and student experience, as an example of institution-specific requirements.
This lesson provides general learning guidance. Always follow the current policy of your school, college or university and check whether that policy has changed.