Anthropic, OpenAI, and Google DeepMind have been engaged in private discussions regarding the creation of an industry-led body for safety standards since July 2026. The initiative gained public attention on September 12, 2026, following an essay by Anthropic CEO Dario Amodei, who proposed a framework for a voluntary slowdown in AI development.
The proposed framework includes granting independent evaluators deep access to frontier AI models for stress testing and establishing global agreements to formalize safety protocols. OpenAI CEO Sam Altman has publicly supported these concepts, specifically highlighting the need to regulate the pace of development.
As of September 14, 2026, the companies have not reached any binding agreements. The process faces significant challenges, including potential antitrust scrutiny from regulators and concerns from smaller industry players that such standards could create barriers to market entry.
