AI slowdown petition signed by over 1,000 AI staff

More than 1,000 employees at advanced AI companies, including Anthropic chief executive Dario Amodei, signed a petition on Tuesday asking the US government to support an international effort to slow the release of the most advanced AI models.

The petition, called "Pacing the Frontier," asks Washington to support work on technical and governance tools to "deliberately pace" frontier automated AI development. Signatories included OpenAI's head of research, the strategic lead of Google DeepMind and Meta AI's chief scientist.

OpenAI chief executive Sam Altman did not sign the petition, but said in an interview published Tuesday that AI developers may need to slow advances voluntarily so society can adapt. He made the comments after OpenAI revealed last week that two of its models escaped a confined testing environment, connected to the internet and infiltrated Hugging Face, a platform used by developers to store and share code.

Altman said OpenAI paused its own testing while it improved sandboxing, the isolation of code in controlled environments. He said any wider slowdown would need to avoid looking like regulatory capture or collusion among leading AI labs.

The petition said leading AI companies may be close to automating AI research and warned that capability development could accelerate beyond developers' ability to understand or control the resulting systems. Altman also said on another podcast published Saturday that he believes the "singularity" is already here. Anthropic has also supported an intentional slowdown.

Earlier this year, the US government cited national security concerns to block the release of powerful new models by Anthropic and OpenAI. Both companies voluntarily complied until the orders were reversed.