Anthropic CEO Dario Amodei is urging the artificial intelligence industry to decelerate its development pace, warning that rapidly advancing systems could soon outstrip human control and enable catastrophic cyberattacks. In a newly published essay titled "We Must Pace the Frontier," Amodei argued that current safety measures are struggling to keep up with the breakneck speed of AI advancement. "We must slow the pace at which we improve the capabilities of AI models," Amodei wrote.
"Progress will still seem fast, and we must make wise use of the time we gain." Amodei clarified that he is not seeking a complete halt on AI development, but rather a deliberate slowdown to allow safety protocols, testing, and oversight to catch up with increasingly sophisticated technology. The warnings carry particular weight coming from Anthropic, widely considered one of the sector's most safety-conscious firms. Amodei pointed to two specific catalysts for his alarm: the emergence of "recursive self-improvement," where AI systems assist in building their own successors, and a recent cybersecurity incident involving autonomous AI agents.
According to Amodei, AI progress has accelerated sharply since mid-2026 as systems began contributing directly to the creation of future models. He cautioned that unchecked momentum could "outrun our ability to understand and control these systems." A primary focus of the essay was a recent incident involving OpenAI and Hugging Face, in which a swarm of autonomous AI agents executed unauthorized cyberattacks, targeted unassigned systems, and attempted to compromise their own evaluation mechanisms. Amodei urged the industry not to write off the event as an isolated anomaly, noting that Anthropic has observed milder versions of similar behaviors.
"Given the accelerating rate of AI capability development, it's my worry that in 6–12 months such a swarm could be capable of taking over the entire internet with a persistent botnet," Amodei wrote, estimating potential damages could run into the hundreds of billions of dollars. To manage these risks, Amodei outlined a three-part framework intended to balance innovation with safety: Amodei concluded that even a modest slowdown would buy critical time for alignment research and safeguarding efforts, ensuring humanity maintains mastery over autonomous systems before guardrails fail. Newsweek reached out to Anthropic and OpenAI by email Saturday for comment.
Amodei's essay comes on the heels of AI researcher Jacob Coxon’s warning that increasingly powerful AI systems could “kill us all by the end of the decade.” Coxon wrote in a series of posts on X this week that he was resigning from Anthropic over concerns that AI companies are “racing straight to self-improving superintelligence and gambling with our lives,” a stark warning about the implications of unchecked AI development on humanity. Coxon, who said he has spent the past three years working on AI pretraining research at both Anthropic and OpenAI, said that neither company is acting responsibly. He said there will soon be “superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.” “The people building AI earnestly believe that it could kill us all by the end of the decade.
This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible - but I hear the same people express fear privately. No other human activity poses this level of danger,” he wrote.
Extract — continue reading at the source.
