Companies as AI Safety Regulators

Powerful AI firms are increasingly acting as the de facto regulators of their own creations: detecting risky capabilities, choosing disclosure channels, assembling private remediation consortia, and deciding who learns what. That private, after‑the‑fact governance can leave the public exposed and bypass democratic accountability. — If firms continue to police catastrophic AI risks themselves, society faces either opaque corporate stewardship or sudden public harms — which argues for rules that mandate independent review, disclosure, and contingency planning.