Anthropic calls for pacing frontier AI development

Anthropic CEO, communicating the company’s stated approach in a public essay.

We must pursue RSI very carefully, if at all.

Source and context

Original text

Dario Amodei — We Must Pace the Frontier (opens in a new tab)Essay sections addressing recursive self-improvement and permanent third-party evaluators

About this source

Direct public essay by Anthropic’s CEO. It sets out an industry-pacing proposal and says Anthropic will give third-party evaluators permanent employee-level access; the statement is a commitment, not evidence that the measures have been implemented.

Before the quotation

Amodei’s essay argues that AI can be beneficial but that accelerating frontier capabilities may outpace the ability to manage safety risks.

After the quotation

The essay proposes permanent external evaluation access and broader coordination among frontier-AI developers and governments.

Mixed or conditional
How this statement is classified

The statement conditions further recursive-self-improvement development on greater caution and describes a safeguard commitment. It does not establish that the measures are implemented or independently prove their effectiveness.

Recorded on
Published here

More from this case

Joe Benton

Frontier AI companies are racing to build AI systems that can recursively self-improve.
Read statement

OpenAI

OpenAI declined to comment when approached by CNBC
Read statement
Read the case