Rights & Ethics
Non-discrimination by design
Proactive design and governance to prevent discriminatory outputs.
Definition
An approach requiring proactive design, testing, documentation, and governance measures throughout an AI system's lifecycle to prevent discriminatory outputs. It encompasses data governance, fairness-aware model selection, feature review, human oversight, testing, monitoring, and remediation measures.
Related Terms
Equitable AI
AI that produces fair outcomes and equitable benefits for disabled people....
Justice/Fairness
Measures to prevent discriminatory or biased outcomes....
Human-centered AI
AI design and deployment that prioritizes human rights and inclusion....
Inclusive AI
AI designed to serve all segments of society equitably....
Algorithmic fairness
Ensuring AI does not discriminate and promotes equity....