Anthropic CEO Outlines Three Steps to Slow Down AI Development

Dario Amodei, CEO of the artificial intelligence company Anthropic, has revealed three steps that could serve as a strategy to slow down the global development of AI. This disclosure comes at a critical time, as growing concerns arise regarding AI-generated content, even as development continues to accelerate at high speeds.

Citing a report from Engadges on Friday (11/9), Amodei proposed a three-level approach, noting that Anthropic has already implemented the first stage. This initial step is crucial for evaluating whether AI model training aligns with the objectives of slowing development and reporting incidents.

The second step involves requiring global AI companies to establish “general safety standards” with government assistance to limit the pace of uncontrolled AI progress. “The steps I am proposing to move forward at a safe speed will not be easy,” Amodei wrote. “But I believe we owe it to humanity to try.”

According to Amodei, the risks of current rapid AI development include “losing control over AI systems, the misuse of AI for cyberattacks and bioterrorism, as well as serious economic disruption.” Anthropic itself recently discovered that some scientists used Claude to create instances of “biological misuse.”

Since the beginning of the year, Amodei’s company has been calling for a slowdown in AI development. Much of this perspective on slowing innovation is reflected in Amodei’s recent blog posts. He suggests that recent incidents serve as further evidence that slowing AI development is necessary to preserve humanity.