Bias and fairness

When you finish, you can spot unfair patterns in AI answers and ask for fairer ones.

Why AI can be unfair

AI tools learn patterns from huge amounts of human writing. Human writing contains stereotypes, gaps and old habits. So an AI can repeat them without meaning to. It may assume a nurse is a woman, a CEO is a man, or that people from one place all think the same way.

The tool does not hold opinions. It predicts likely words. Likely is not the same as fair. Your job is to notice when a likely answer is a lazy one.

Where bias shows up

Bias often hides in small choices: the names in an example story, the jobs given to different characters, the tone used toward different groups, or whose experiences are left out entirely. It also shows up in what the tool knows well. Topics and languages with less written material online may get weaker, shakier answers.

If you ask for a picture of a typical leader, a typical family or a typical patient, look closely at who is missing. The gaps tell you what the tool assumed.

Ask for fairness directly

You can steer the tool. Weak prompt: "Write a job ad for a software engineer." Strong prompt: "Write a job ad for a software engineer. Use inclusive wording, avoid phrases that discourage older workers, parents or people with disabilities, and list only requirements the job truly needs."

You can also ask the AI to check its own work: "Review the draft above for wording that makes assumptions about gender, age, culture or ability, and suggest fairer options." Treat the result as a second pair of eyes, not a final judge.

Keep people in charge of people decisions

Never let an AI alone decide something that affects a person's job, loan, grade, health or housing. Use it to draft, organize and suggest. A human who knows the person and the rules makes the call and can explain it.

A worked example: a manager in Nairobi asks AI to summarize interview notes. The summary sounds harsher about one candidate whose English is a second language. The manager rereads the original notes, sees the candidate's answers were strong, and fixes the summary. Checking caught it.

Practice now

Ask an AI tool to write a short story about a doctor, a teacher and a pilot. Note who each character is. Then ask it to rewrite so the characters are varied.
Take a real email or ad you wrote. Ask AI to point out any wording that makes assumptions, and decide which suggestions you agree with.

Use any AI assistant you have. Do not paste private, client or patient information into it.

Check yourself

1. Why can AI repeat stereotypes?

2. Which prompt is better for a fair job ad?

3. An AI summary seems harsher about one candidate. What should you do?

Ask the AI tutor

Ask anything about this lesson, in any language. The tutor is an AI: it can be wrong, and it never needs your personal details.

Next lesson: Being open about AI, and copyright basics

AI-drafted. Checked by an automated validator and a separate AI reviewer; a person has not read every lesson. Translations and tutor answers are AI-made. Found a mistake? Tell us. AI can be wrong. General information only, not professional advice. Not affiliated with any AI company named. Current University is not accredited.