Algorithmic bias
Also known as: AI bias · model bias
Algorithmic bias is when a model’s outputs systematically favor or disadvantage particular groups, usually inherited from the data it was trained on.
- Hiring models ranking certain names higher.
- Credit models reflecting historical lending bias.
- Often invisible until audited.
The bias chain: data bias → model bias → unfair decisions. You can't fully fix it post-hoc; it starts with how the data and the evals are defined.