AI bias and discrimination incidents
Definition
Bias and discrimination incidents document cases where an AI system produced outcomes that were unfair, inequitable, or discriminatory toward individuals or groups based on characteristics such as race, gender, age, disability, nationality, or other protected attributes.
Included
- Discriminatory decisions in hiring, lending, housing, healthcare, or criminal justice made by AI systems
- AI facial recognition with documented disparate error rates across demographic groups
- AI content moderation that disproportionately affected specific communities
- Algorithmic systems that amplified or perpetuated structural inequalities with documented harm
- AI-generated content that denigrated or stereotyped groups based on protected characteristics
Excluded
- Statistical disparity claims without a documented event or measurable harm
- Theoretical fairness critiques of AI systems that have not caused a specific incident
- Cases where the discriminatory outcome was driven entirely by human policy rather than AI behavior