AI developers should compromise on model capability to guarantee human control

Artificial intelligence research, products, and impact across society.

The frameworks, policies, and standards that guide the responsible development, deployment, and oversight of artificial intelligence systems.

The field of preventing harm from advanced AI systems, spanning alignment, evaluation, and decisions to slow or withhold models whose capabilities outpace their safeguards.

Practices ensuring ethical and accountable AI use. Includes fairness and transparency.