🤖 Case 1: The Hiring Bot

What’s happening
A company uses an AI system to scan résumés and rank applicants.

What’s problematic
The bot often rejects women with career gaps (like time taken off for childcare). Why? Because it learned from biased historical data that favored men with straight career paths. The result: hidden bias baked into the hiring process.

One improvement idea
Run fairness checks with diverse test résumés, add explainability so recruiters see why a profile was rejected, and let applicants challenge AI decisions.

👀 Case 2: The School Proctoring AI

What’s happening
An AI tool monitors students during online exams, flagging “suspicious” eye movement.

What’s problematic
Students who are neurodivergent or anxious may naturally move differently, and the AI often mislabels them as cheaters. That’s not only stressful, but also unfair.

One improvement idea
Instead of relying only on eye movement, combine multiple signals (like unusual answer patterns + keystrokes) and always include human review before penalizing a student.

✨ Final Note

AI isn’t evil, but it learns patterns, not people. Without fairness, transparency, and accountability, it risks acting like a judge rather than a helper.

The fix? Treat AI like a rookie detective: give it oversight, guidance, and ethical guardrails. Only then can it be the trusted sidekick we all need.