Anthropic, OpenAI and Google Pursue Industry-Led AI Oversight
A voluntary agreement in September 2026 called for companies to put stronger controls in place. The U.S. administration supports the wider effort, but the industry does not yet have one set of rules that applies to all companies.
Anthropic, OpenAI and Google are moving forward with an effort to let companies oversee advanced AI themselves. The U.S. administration supports this effort, but there is no single set of rules for the whole industry.
The effort follows a voluntary AI safety agreement from September 2026. It called on developers of advanced AI models to use stronger internal controls and better ways to manage risks. The new approach does not set rules that apply to every company in the industry.
Under the proposed plan, companies would take the lead in overseeing their AI systems without direct government involvement. The agreement called for stronger internal controls. Companies may also work together on shared safety practices, but they have not made a firm commitment to do so.
This difference matters for U.S. technology policy. Support from the administration has not led to one common process for overseeing the companies involved. For now, each developer is focusing on the steps it chooses to take.
Riza Dagoc
Riza Dagoc reports on rules and regulations, investing, and the digital-asset business for TokenPost. Send corrections or tips to [email protected].