OpenAI, Anthropic, Google Collaborate on AI Safety Standards

OpenAI, Anthropic, and Google DeepMind have been collaborating for several weeks on artificial intelligence safety standards, according to OpenAI's global policy chief Chris Lehane. The effort aims to establish a private-sector-led standards body for testing and auditing advanced AI models, a proposal that originated from Google DeepMind founder Demis Hassabis. Lehane confirmed the talks during a Washington briefing on Tuesday, stating that OpenAI does not believe an antitrust waiver is needed for the three companies to coordinate on safety matters. He cited precedents like airline safety cooperation, where companies share sensitive information under existing antitrust law. The initiative faces skepticism from U.S. officials. Federal Trade Commission Chairman Andrew Ferguson said Tuesday that policymakers should be cautious about AI companies seeking antitrust exemptions while also requesting new regulations. Ferguson framed his remarks as a personal view rather than official policy. OpenAI is supporting bipartisan legislation to address AI risks.