Trust & Safety

Bias

Also called: AI

AI models reflecting and amplifying biases present in their training data. If training data underrepresents certain groups or contains historical discrimination, the model may reproduce those patterns. Bias is not unique to AI, but AI applies biased judgments at scale and speed. Bias auditing is especially critical for AI used in hiring, lending, healthcare, or legal contexts.

Lessons that cover this

This definition is part of a free, structured course on how AI actually works.

Start learning