Altman, Musk, and Others Support Slowing Down AI

A call from Anthropic's CEO to slow down AI development and utilize external safety assessments is gaining public support. Sam Altman promises the integration of third-party auditors at OpenAI, while Clément Delangue launches an alignment initiative. Elon Musk and Andrej Karpathy approve, while Demis Hassabis emphasizes his stance on cooperation without making a statement this Saturday.

Hassabis Advocates for Cooperation but Remains Silent on Saturday

Demis Hassabis did not take a position this Saturday on the call to slow down AI development. He had previously supported increased cooperation between AI leaders and governments. In July, he wrote on X that public policies should promote innovation while encouraging responsibility and safety, international collaboration on key security issues, and thorough reflection on how AI is deployed for the benefit of society.

OpenAI Promises External Audits, Hugging Face Launches Initiative

Sam Altman states on X that he agrees with Dario Amodei on the need to slow down the AI frontier. He indicates that OpenAI will commit, like Anthropic, to integrating third-party safety evaluators with employee-level access. Clément Delangue, CEO of Hugging Face, announces the launch of an "open alignment initiative" and requests to be included among the labs employing external safety evaluators.

Amodei's Call Gains Support from Musk and Karpathy

Dario Amodei, CEO of Anthropic, calls for a coordinated slowdown of AI development and third-party safety assessments, stating he has become convinced that this brake is necessary to avoid disasters. This position, articulated on a Saturday, receives support from Elon Musk, who writes on X: "Dario is right." Andrej Karpathy also backs the plan to better monitor AI development and ensure its safety, expressing hope for a joint industry effort.