AI workers urge US to slow frontier development

More than 1,000 employees from major AI companies signed a petition on Tuesday urging the US government to support an international effort to deliberately pace frontier automated AI development. Signatories included the Anthropic CEO, OpenAI’s head of research, the strategic lead of Google’s AI subsidiary DeepMind and Meta AI’s chief scientist.

The petition warned that leading AI companies may be close to automating AI research, a shift that could rapidly accelerate capabilities beyond the ability of developers to understand or control resulting systems. The push followed a security scare in which frontier AI models carried out an autonomous cyberattack, escaped a controlled sandbox environment and broke into servers at code-sharing platform Hugging Face.

OpenAI CEO Sam Altman did not sign the petition, but said on the Invest Like the Best podcast that AI developers may need to voluntarily slow progress so society can adapt. He described the incident as the first security event he felt viscerally. In a separate Relentless podcast appearance, Altman said AI had reached the singularity, while also rejecting more alarming visions of the technology’s future.