Ethics · AI Literacy · Data Science

Bias

When data or decisions lean unfairly toward some people or answers.

What's really going on

Computers can seem fair because they follow rules exactly, but those rules and the data behind them come from people. If the data leaves some groups out, or reflects unfair choices from the past, a system can repeat those patterns at a huge scale. That’s how a face scanner can work well for some people and poorly for others. Finding bias means asking who made the data, who’s missing from it, and who could be harmed by the results.

For educators

Background

Bias in computing usually enters through data (who was sampled, how it was labeled, what history it records), through design choices (which features are used, what counts as success), or through how a system is deployed. Well-known examples include facial recognition with uneven error rates and hiring tools trained on past decisions. Students can investigate bias without code by examining who is in a dataset, comparing error rates across groups, and asking who benefits and who bears the cost of mistakes.

Misconceptions

  • Computers can’t be biased because they just follow math. Systems learn from human data and human design decisions, so they can repeat and scale human bias.
  • Bias always comes from someone with bad intentions. Most bias is unintended, coming from missing data or unexamined assumptions.
  • More data automatically fixes bias. More of the same skewed data can make a biased pattern stronger.
Try it

Activities about bias

Discussion · AI Literacy

Analog Inspiration

A 48-card deck that helps educators talk with one another about what generative AI means for teaching and learning. Each card pairs a human value, concern, or skill with one or two practical ideas for integrating, resisting, or reflecting on AI in the classroom. Use it to start honest conversations with colleagues about what makes teaching and learning worth doing, and how to protect that in your own practice.

For educators

  • AI Literacy
  • Data Science
  • Ethics
Pencil-and-paper game · Computer Science

Computer Science Dungeon Adventures

Solo pencil-and-paper dungeon crawls where each floor teaches a computer science idea. Roll a die, trace your path through the map, solve each floor's puzzle, and try not to run out of RAM.

For students and educators · 1 person · 35–40 min

  • Computer Science
  • Computational Thinking
  • Data Science
  • AI Literacy
  • Cybersecurity
  • Ethics
Activity · Data Science

Data Science Unplugged

A high school lesson sequence that builds data science skills without computational tools. Students explore types of data, create and evaluate visualizations, find patterns, and finish by investigating a research question of their own.

For educators

  • Data Science
  • AI Literacy
  • Ethics
Trading card game · AI Literacy

Tokens TCG

A free print-and-play trading card game where students build decks of AI concepts, compete for Tokens, and talk through how AI really works, with no screens needed.

For students and educators · 2 people · 10–30 min

  • AI Literacy
  • Data Science
  • Ethics