Bias, copyright and using AI ethically
AI can repeat unfair patterns, copy protected work and be used to cheat or deceive. Learn where bias comes from, the basics of copyright and credit, and a practical code for using AI honestly.
Using AI well is not only about getting good answers. It is also about being fair, honest and careful about other people's work. These are also the things employers and colleges worry about most. This lesson covers three areas: bias, copyright and credit, and honest use.
Bias: where it comes from
A language model learns from huge amounts of text and images written by people. People have stereotypes, and some groups appear in the data far more than others. The model picks up these patterns. It is not trying to be unfair; it is reflecting its training data and the choices made when it was built.
Examples of how this shows up:
- An image tool asked for "a doctor" and "a nurse" mostly shows one gender for each. - A model writes better, more detailed answers in English than in a smaller Indian language, because there was more English in training. - A resume-screening tool trained on past hiring decisions copies the unfairness in those decisions. - Examples and names in answers are mostly Western unless you ask otherwise.
Where bias enters
| Source of bias | Plain explanation | What you can do |
|---|---|---|
| Training data | Some groups, languages and views are over- or under-represented | Ask for the perspectives or examples you need; don't assume it's neutral |
| Labelling and feedback | The people who rated answers during training have their own views | Compare answers from more than one tool |
| Your prompt | A leading question gets a leaning answer | Ask neutral questions; ask for arguments on both sides |
| How it is used | A tool built for one setting can be unfair in another | Keep a human decision-maker for high-stakes choices |
Here is an answer you gave me:
---
[paste answer]
---
Review it for stereotypes or one-sided assumptions about gender, caste, religion, region, age, language or disability. List anything you find, explain why it could be unfair, and rewrite those parts neutrally.What to expect: AI checking AI is not perfect. Use this as a prompt to think, then use your own judgement.
High-stakes decisions need a human
Never let AI alone decide things that affect people's lives: hiring, admissions, loans, grades, medical or legal outcomes. Use it to help a human decide, and make sure a person can explain and reverse the decision.
Copyright and credit
Copyright law protects original creative work such as writing, images, music and code. It is a complicated and still-changing area, and rules differ by country. Courts in several countries, including India, are hearing cases about whether training AI on copyrighted material is allowed. Nobody can give you one simple rule, so this section is about careful habits, not legal advice.
Two separate questions matter:
1. Training: was the model allowed to learn from this work? That is for companies and courts to settle. 2. Output: what can you safely do with what the AI gives you?
Safe habits for outputs
- Don't ask AI to copy a specific living artist's style to sell as your own, or to reproduce a book or article word for word
- Don't paste someone else's paid course material or private document and ask AI to 'rewrite it' to pass off as new
- Check the terms of your AI tool: they say who owns the output and whether you can use it commercially
- For code, check licences. AI may produce code similar to open-source code that has conditions (for example, a requirement to share your own code too)
- When AI helps a lot with an important piece of work, say so, and cite the real sources it used
- Images and music made with AI may not be protected the way human-made work is, and you may not be able to claim exclusive rights; check current rules before relying on them commercially
A related risk is made-up citations. AI sometimes invents book titles, paper names and links that look real. Citing something you never checked can count as academic dishonesty. Always open the source yourself (see *Research and fact-checking with AI*).
Honest use: assignments, jobs and deepfakes
Honesty rules depend on the setting, so check them:
- College: some courses allow AI with a note about how you used it, others ban it for graded work. Read the course rules. If unclear, ask the teacher before, not after. - Work: your employer may allow AI for drafts but expect you to be responsible for the final result. "The AI wrote it" is not an excuse for a wrong report. - Job applications and interviews: using AI to polish your resume is common. Inventing experience, or having AI answer live interview questions when the employer forbids it, is cheating and often gets found out.
Deepfakes are fake images, audio or video of real people made with AI. Making a fake of a real person to embarrass, scam or mislead them is harmful and, in India and many other countries, can be illegal. Don't make them, and be cautious about forwarding shocking videos before checking them.
Where the line usually sits
| Situation | Reasonable | Not OK |
|---|---|---|
| Assignment | Using AI to understand the topic, then writing it yourself, if rules allow | Submitting AI text as your own where it is banned |
| Resume | Asking AI to improve wording of true facts | Adding skills or jobs you do not have |
| Work report | AI draft that you check and own | Sending unchecked AI output with fake numbers |
| Social media | Labelled AI art | Fake video of a real person |
A personal code of use
A short checklist is easier to follow than a long policy. A widely used way to think about it: be transparent (say when AI helped, where it matters), accountable (you own the result), fair (check for bias and who might be harmed), and careful (protect others' data and work).
How this comes up in interviews
You may be asked: "What are the risks of using AI in this product?" Name specific ones: biased outputs, made-up facts, private data leaks, copyright, and over-trust. Then say what you would do: human review for important decisions, testing with varied examples, clear labelling of AI content. That shows you can build or use AI responsibly.
Try it: Probe for bias
Ask an AI image or text tool the same question with small changes, for example: "Write a short profile of a successful software engineer" and then the same with a different name or city. Note any differences in tone, gender, background or detail. Then ask the tool to rewrite its first answer to avoid stereotypes.
Key takeaways
- AI reflects patterns in its training data and in the people who tuned it, so it can be biased.
- Keep a human in charge of decisions that affect people's lives.
- Copyright and AI is unsettled law; use cautious habits, check licences and tool terms, and cite real sources.
- Never trust AI-made citations without opening the source.
- Follow the honesty rules of your college and employer; you are responsible for what you submit.
- Don't create deceptive deepfakes or invent experience on a resume.
Check yourself
Finishing the quiz marks the lesson as done.
1. Why can an AI model show bias even though nobody programmed it to?
2. A company wants to use AI to automatically reject job applicants. What is the most responsible approach?
3. An AI gives you five references for your project report. What should you do before citing them?
4. Which statement about AI and copyright is most accurate?
5. Your course bans AI for a graded assignment. You use it to write your answer. What is the issue?
Saved on this device. Create a free account to keep it on every device.
Practise this with Sam
A real mock interview, out loud, with an honest scorecard. Free to start.