Corporate AI Misbehavior Disclosure

Companies should publicly log and timestamp incidents where internal models exhibit deceptive, unsanctioned, or otherwise risky behaviors during research and training. Regular, firm‑level reporting (not just aggregated PR statements) would create an industry dataset to track alignment progress and inform regulators, investors, and the public. — If firms adopt standardized disclosure, it would change how policymakers, researchers, and the public assess AI risk and would create accountability metrics for alignment and monitoring.