(VIDEO) Anthropic CEO Dario Amodei Tells CNN’s Anderson Cooper He Fears AI Could Threaten Humanity This Decade

Anthropic CEO Dario Amodei told CNN’s Anderson Cooper in an exclusive interview that he largely agrees with a former company researcher who resigned this year warning that the people building artificial intelligence privately believe the technology could kill humanity by the end of the decade, while cautioning against reducing the risk to a single fixed number.

The interview, which aired on “Anderson Cooper 360,” came the same weekend Amodei published a lengthy essay on his personal website calling for the AI industry to deliberately slow the pace of its development, a position that has since drawn public support from rival executives Sam Altman of OpenAI and Elon Musk of xAI.

A pointed question about extinction risk

Cooper’s exclusive interview centered on comments made by Jacob Coxon, a former Anthropic researcher who resigned earlier this year and publicly warned that “the people building AI earnestly believe that it could kill us all by the end of the decade.” A current Anthropic alignment researcher, Evan Hubinger, has separately estimated a greater than 10% chance of AI causing human extinction within the next 10 years.

Asked directly whether he shares that assessment and what probability he would assign to it, Amodei responded candidly rather than dismissively. “I agree with Jacob much more than I disagree with him,” Amodei said. He noted that Coxon had specifically praised Anthropic as the most responsible actor in the industry while directing his broader criticism at the pace of development across AI companies generally, rather than at Anthropic specifically.

Why Amodei resists putting a number on the risk

Amodei pushed back against the idea of reducing the danger to a single statistical probability, arguing that doing so misrepresents the nature of the risk itself. He said assigning a fixed percentage chance to AI causing catastrophe makes the situation sound like “a roll of the dice,” when in reality the outcome remains contingent on choices the industry and governments make going forward.

“If we take the right paths, then the chance of something going wrong is very low,” Amodei said. “If we take the wrong path, then the chance of something going wrong could be even higher than that.” He framed the central challenge facing the industry as one of agency, arguing that the priority should be steering AI development toward safer paths through industry cooperation and, eventually, broader coordination at a global level.

Concerns about private control of powerful AI

Beyond the extinction-risk question, Amodei also addressed his discomfort with the concentration of advanced AI capabilities in the hands of private companies. Speaking with Cooper, Amodei said he remains uneasy about the current arrangement in which frontier AI development sits largely within a small number of privately held corporations, arguing that there needs to be a better balance struck between government oversight and private-sector control as the technology continues to advance. He characterized the technology as something that ultimately “belongs to all of us,” rather than something that should remain the exclusive purview of the handful of companies currently building it.

A weekend essay calling for a slower pace

Amodei’s CNN interview followed closely on the heels of a roughly 3,800-word essay he published early Saturday, in which he laid out a formal case for the industry to pump the brakes on the speed of AI capability improvements. “We must slow the pace at which we improve the capabilities of AI models,” Amodei wrote in the essay. “Progress will still seem fast, and we must make wise use of the time we gain.”

In the essay, Amodei acknowledged that humanity could ultimately lose control of advanced AI systems altogether, and he warned that the technology could be misused to facilitate “cyberattacks and bioterrorism, and serious economic disruption.” Amodei, who has built Anthropic into one of the world’s leading AI developers and created its chatbot Claude, has long positioned himself as a strong advocate for AI’s continued advancement, even as he has simultaneously faced accusations of “doomerism” over his repeated public warnings about the technology’s risks.

Rival executives voice rare agreement

The essay drew an unusual show of public support from two of Amodei’s most direct competitors. OpenAI CEO Sam Altman responded on X, writing that he agreed with Amodei’s proposal for independent oversight of AI companies’ safety practices. “Committing to having independent evaluators with employee-like access is a great idea, and we will do the same,” Altman wrote.

Musk, whose company xAI develops the competing chatbot Grok, offered a brief but pointed endorsement of his own, posting simply, “Dario is right.”

A regulatory vacuum in Washington

Amodei’s calls for a coordinated slowdown come against the backdrop of continued inaction in Washington on comprehensive AI regulation. Congress has debated legislation addressing AI safety standards but has not passed any overarching framework, and the White House has yet to reach internal consensus on which federal agency or office should ultimately bear responsibility for overseeing the rapidly advancing technology, according to prior CNN reporting.

A broader industry reckoning

Cooper’s interview also touched on recent incidents involving AI agents behaving unpredictably, including episodes involving the companies Hugging Face and OpenAI, which Amodei cited as evidence supporting the need for more caution industrywide. Taken together, Amodei’s CNN interview and his accompanying essay mark one of the most direct public acknowledgments yet from a sitting AI company CEO that the technology’s current trajectory carries genuine, near-term risks that the industry itself may not be equipped to manage without significant changes to how quickly new capabilities are deployed.