AI Enforcement Liability

The Federal Trade Commission is proposing to treat intentional AI output steering that conflicts with consumers’ reasonable expectations of accuracy or objectivity as a potential deceptive practice under Section 5 of the FTC Act. The proposal distinguishes deliberate design choices from ordinary model errors and emphasizes that prominent disclosures may be needed when systems pursue objectives users would not expect. The broader material also shows FTC enforcement of unsupported AI capability and consent claims, while courts and legal institutions are emphasizing tool vetting, privacy protection, and human verification rather than delegating judgment to AI.

U.S. lawmakers and regulators are advancing different approaches to AI oversight, with Congress considering targeted incident-reporting requirements while states pursue broader rules for frontier model audits, employment tools, AI companions, and synthetic media. Illinois, Connecticut, and California illustrate the growing reach of state obligations, while federal lawmakers and industry groups remain divided over whether national standards should replace or preempt state rules. The resulting debate could determine how companies report safety failures, demonstrate compliance, label AI-generated content, and manage liability across jurisdictions.