THE CRUNCH
Sam Altman is urging a more measured approach to artificial intelligence development, proposing a uniform national safety framework while insisting that progress will not slow down. He says OpenAI will set explicit safety protocols before training runs that could bring major capability jumps, and hopes other companies will adopt similar measures. Altman argues that pacing is necessary to ensure capabilities do not go
ahead of alignment and monitoring, warning that no amount of competitive pressure should justify recklessness. This stance has support from other industry leaders, including Anthropic's Dario Amodei and Google DeepMind's Demis Hassabis, who have backed better-controlled development. According to The Information, OpenAI, Anthropic, and Google have been discussing self-regulation through an independent oversight body, though smaller players like Cohere have raised concerns that such efforts could turn into a cartel.
Altman acknowledges that international coordination requires U.S. government support, even as he advises the industry to move forward without waiting for lawmakers. He frames 'pacing' not as a halt, but as a deliberate approach to ensure safety while maintaining a competitive edge. The push for self-regulation comes amid broader industry discussions about how to manage the risks of increasingly powerful AI systems.