Anthropic researchers warn out-of-control AI could destroy humans

There is a peculiar irony in the fact that the people most worried about AI are the very ones building it. When Jacob Coxon, who has worked at both Anthropic and OpenAI, steps forward to say neither company is acting responsibly, he is not speaking from the outside looking in—he is speaking from the engine room. That gives his warning a weight that purely theoretical doomsaying lacks. He has seen the trajectory from the inside, and he does not like where it is heading. The deeper story here is about incentives. AI companies operate in a hyper-competitive market where being second can mean being obsolete. Even when executives sincerely believe in safety, the pressure to ship the next breakthrough model is immense. Anthropic, ironically, was founded partly as a safety-first alternative to OpenAI—yet now its own researchers are saying that mission has been compromised. This is not a story about evil corporations; it is a story about structural forces that push even well-intentioned organizations toward speed over caution. There is, of course, another side worth considering. Critics of the "AI extinction" framing argue that these warnings are overhyped, that the technology is nowhere near superintelligence, and that focusing on far-future catastrophe distracts from concrete, present-day harms like bias, misinformation, and job displacement. Some also note that former employees have financial incentives—book deals, speaking fees, new ventures—to amplify dramatic narratives. These are fair points, and they deserve a hearing alongside the alarm bells. What makes this moment genuinely different is the unprecedented nature of the technology itself. Previous existential threats—nuclear weapons, climate change—were understood well enough that experts could model the risks with some confidence. With AI, even the experts admit they are operating in profound uncertainty. When the people closest to the technology say they cannot guarantee its safety, that uncertainty is not a reason for panic, but it is a reason for humility—and for policymakers to stop treating this as a purely private-sector matter. The Hill's reporting on this story (https://thehill.com/policy/technology/6078907-anthropic-researchers-warn-ai-extinction/) captures the tension at the heart of the debate. Whether Coxon is right or wrong, the fact that researchers feel compelled to go public suggests the internal safeguards many assumed were in place may not be as robust as advertised. In a democracy, that is exactly the kind of information citizens need—not to be frightened into inaction, but to make informed choices about how much risk we are willing to accept in exchange for technological progress.