OpenAI, Anthropic face new oversight as New York’s AI safety law takes effect

New York's AI safety law adds a new compliance layer for major model developers, with direct implications for teams building on systems from OpenAI, Anthropic, and other frontier labs. Covered companies must register with the state, disclose their safety practices, and report critical incidents. Failure to comply can result in fines. For practitioners, the immediate impact is likely to appear in vendor documentation, incident-notification terms, and procurement reviews rather than model APIs themselves. The law may give enterprise customers more visibility into how providers test and govern high-capability models, while forcing developers to formalize reporting processes for serious failures. The key details now are how New York defines covered companies and critical incidents, and how aggressively regulators enforce the requirements.