Results for "regulatory compliance"

AI giving legal advice without authorization.

Requirement to reveal AI usage in legal decisions.

Ensuring models comply with lending fairness laws.

Simulating adverse scenarios.

Sudden extreme market drop.

Rate at which AI capabilities improve.

Regulating access to large-scale compute.

Automated detection/prevention of disallowed outputs (toxicity, self-harm, illegal instruction, etc.).

Techniques to understand model decisions (global or local), important in high-stakes and regulated settings.

Central system to store model versions, metadata, approvals, and deployment state.

A discipline ensuring AI systems are fair, safe, transparent, privacy-preserving, and accountable throughout lifecycle.

Extracting system prompts or hidden instructions.

Privacy risk analysis under GDPR-like laws.

Central log of AI-related risks.

Legal right to fair treatment.

Restricting distribution of powerful models.