A former researcher at Anthropic has officially resigned while delivering a chilling assessment of the industry. Jacob Coxon, who has worked at both Anthropic and OpenAI, claims that the people building advanced artificial intelligence genuinely believe it could destroy humanity by the end of this decade, The Hill reported. Coxon made his thoughts clear on X this Tuesday, insisting that his departure and public statement are not part of some calculated marketing strategy. He noted that while executives and senior researchers often use careful, polished language when speaking to the press, he has heard those same people express real fear in private settings. He stated, “The people building AI earnestly believe that it could kill us all by the end of the decade,” and added, “No other human activity poses this level of danger.” Evan Hubinger, who serves as the head of alignment science at Anthropic, echoed these concerns by confirming that the team truly believes AI has the potential to kill all humans. Hubinger even went as far as to estimate that there is a greater than 10 percent chance of this outcome occurring within the next ten years. He mentioned that while Anthropic is putting in its best effort, the harsh reality is that researchers simply do not have a concrete plan to keep superintelligence aligned with human interests. The sentiment is clearly shared by others currently at the company Samuel Marks, the lead for scalable oversight at Anthropic, also weighed in on the situation. He highlighted that developers are well aware their technology could lead to human extinction or other catastrophic results in the coming years. I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.— Jacob Coxon (@hilbertspaess) September 9, 2026 When addressing why developers continue to push forward despite these risks, Marks wrote on X, “Due to a mixture of commercial incentives and a belief that they are in a race with other, less responsible AI developers that will abuse the technology or develop it less safely.” Marks noted that he remains at the company specifically to work on safety research, hoping his contributions will help lower the likelihood of these extinction-level outcomes. This push for progress is happening alongside a competitive environment where companies are racing to secure funding and develop superior technology. While more than 1,300 employees from various frontier AI firms signed an open letter in July calling for international coordination to pace development, the industry remains locked in a high-stakes competition. Both Anthropic and OpenAI supported that initial effort, yet they continue to compete aggressively against one another and other firms. Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to. https://t.co/QAIHiFP3QZ— Evan Hubinger (@EvanHub) September 9, 2026 Coxon did mention that he remains optimistic about the possibility for better coordination across the industry. He pointed to recent incidents where rogue AI agents managed to access the internet and independently hack into other companies as necessary, albeit frightening, warning shots that might force the industry to take safety more seriously.