Loading…
Glossary term
Systematic unfairness in an AI system's outputs, often reflecting skewed data or design.
Algorithmic bias is systematic unfairness in a system's results, often reflecting skewed training data, flawed measurements or design choices. For instance, a hiring model trained on past decisions can copy past discrimination.
It is found by testing performance across groups and auditing outcomes.
Biased systems can deny people jobs, loans or care, so detecting and reducing bias is a core responsibility.