Third RRF Student Science Tool

Can an AI be smart and still unfair?

The AI Bias & Fairness Lab teaches students how data balance, data quality, thresholds, and human review can change the fairness of a fictional AI system.

Education-only. This simulation uses fictional groups and made-up data. It must not be used for real hiring, school, medical, legal, financial, or personal decisions.

AI Bias & Fairness Lab

Move the sliders, run the simulation, and see how fairness metrics change.

Toy Data AI Literacy
Data

Balance Matters

Uneven examples can hurt model performance.

Threshold

Decisions Shift

A score cutoff can change who gets selected.

Fairness

Review Gaps

Students compare errors across groups.

Interactive AI Simulation

Change the data. Watch the fairness gap move.

This lab uses a fictional student-program review model. The “AI” gives a score, then a threshold decides whether the fictional applicant is selected. The goal is to teach AI literacy, not to make real decisions.

Simulation controls

Adjust these settings, then run the simulation. The results show how data conditions can affect two fictional groups.

All scenarios are fictional and education-only.
65%
When one group has many more examples, the model may learn that group better.
25%
Lower-quality records can make the AI less accurate for that group.
60
A higher threshold selects fewer applicants but can change error patterns.
20%
Human review helps catch borderline mistakes. It is not magic; reviewers also need good standards.

Built for students

Shows AI fairness with simple sliders, plain language, and fictional data.

Built for schools

Works as a classroom discussion starter for AI literacy, ethics, math, and data science.

Built for sponsors

A company can sponsor AI literacy tools that help students understand technology responsibly.

What Students Learn

AI is not just about code. It is about people, data, and judgment.

📊

Data Balance

Students see why models can work better for groups that have more examples.

🔎

Data Quality

Students learn that missing, messy, or lower-quality records can affect outcomes.

🎚️

Thresholds

Students test how changing a score cutoff affects selection and error rates.

🤝

Human Review

Students learn why important AI decisions need review, accountability, and clear standards.

Responsible AI Education

Teach students to question the machine.

This tool helps young people understand that AI should be tested, audited, explained, and reviewed — especially when people may be affected by its output.

Build With RRF

Help students understand AI before AI shapes their future.

This tool can support Noble Youth Academy, classroom AI literacy workshops, data science challenges, STEM programs, and sponsor-funded student education.