Anthropic CEO Dario Amodei calls for slower AI progress and more oversight

In short: Anthropic CEO Dario Amodei is again warning about worst case AI risks, and he is urging slower development with stronger outside oversight.

What's going on

Dario Amodei, the co-founder and CEO of Anthropic, has a reputation for sounding alarms about artificial intelligence. Some people call him a “doomer,” meaning someone who expects disaster. Others argue he is being responsible because he has spent years working on advanced AI systems and has seen how fast they are improving.

Amodei’s views were shaped in part by his time at OpenAI. The reporting says he grew worried that AI “capabilities,” what the systems can do, were moving faster than safety work like testing and controls. That disagreement helped lead to Anthropic’s founding in 2021 by Amodei and other former OpenAI employees, with a stated focus on safety.

In a September 2026 essay, Amodei warned that more capable AI “agents” could increase the risk of harm. AI agents are systems that can take actions for a user, like a very fast assistant that can click buttons and run tasks on its own. He raised concerns about cyberattacks, bioterrorism, and major economic disruption, and he called for independent evaluators who get deep access to AI companies, shared safety standards, and coordination between countries.

What to watch

The argument over Amodei’s stance is unlikely to fade. Supporters say outside scrutiny and common rules are needed. Critics question whether companies building powerful AI can truly police themselves, and whether scary warnings could push regulations that favor big firms over smaller competitors.

Source: NYTimes