Anthropic, OpenAI and Google Pursue Industry-Led AI Oversight

Anthropic, OpenAI and Google are advancing a company-led effort to oversee advanced AI, with U.S. administration support but no uniform set of industry requirements in place.

The effort follows a voluntary AI safety accord from September 2026 that called for stronger internal controls and risk management among developers of advanced AI models. The emerging framework does not set requirements that apply across the industry.

Under the proposed approach, companies would take the lead in managing oversight without direct government intervention. The accord called for stronger internal controls, while coordination on shared safety practices remains a possibility rather than an established commitment.

The distinction matters for U.S. technology policy: administration support has not produced a common oversight process for the companies involved. For now, the effort centers on measures adopted by individual developers.