Tech giants explore joint AI safety standards initiative

Anthropic, OpenAI, and Google are reportedly advancing plans to form an AI industry standards organization, with discussions dating back to July. The working group, composed of mid-level representatives from the three companies, has held regular meetings, including one last week. Anthropic's CEO, Dario Amodei, recently called for slowing AI development and establishing independent oversight, proposing voluntary safety standards alongside government regulation. OpenAI's Sam Altman supports industry-led standards but opposes government involvement, citing lack of consensus. Demis Hassabis of Google DeepMind endorsed the idea but noted the need for refinement, while Elon Musk also expressed support. The White House had previously proposed a regulatory body but failed to secure enough support, leading to a pause in related efforts. Critics, including Cohere's Aiden Gomez and former White House AI director David Sacks, raised concerns about potential antitrust issues and questioned the altruism of slowing development. Meta's stance remains unclear, with its Chief AI Officer Alexander Wang offering only a general statement on alignment.