A former Anthropic researcher warned that the race to build increasingly powerful artificial intelligence could eventually produce technology that humans can no longer control. Jacob Coxon, who recently resigned from the tech giant over such concerns, told MS NOW’s Ari Melber that without stronger guardrails and regulation, it could soon be “lights out” for humanity.
Coxon previously left OpenAI for similar reasons, and now says both companies are moving too quickly toward increasingly autonomous AI without adequate safeguards.
During the interview, Coxon outlined a potential future in which AI could outthink humans: “Imagine the computer is vastly more intelligent than you. The AI can think faster than you. It can copy itself. It maybe has a thousand copies of itself, all working in tandem. And now, imagine that the AI, for whatever reason, decides it doesn’t want you around anymore. It wants to be in control.”
“As soon as the AI makes that decision, or as soon as the AI accidentally stumbles into thinking that it has to make that decision, it’s basically lights out if it’s smart enough,” he warned.
Coxon stressed that while today’s AI systems are not yet capable of those extreme scenarios, they have already caused real-world damage, citing a recent hack that utilized AI agents from OpenAI.
That danger, according to Coxon, only increases as the systems become more autonomous and connected to the physical world. While it may seem harmless to use AI to control light switches or other devices, “there are a lot of ways that the computer can actuate itself in the real world because basically everything’s connected at this point,” he warned.
Coxon is among many working inside the industry who have called for greater caution as companies race to build increasingly capable AI. He said he supports efforts such as those outlined in a July 2026 open letter, “Pacing the Frontier,” which calls for regulation and a more measured approach to AI development.
“I would like there to be some form of slowdown,” he told Melber.
The former Anthropic employee acknowledged that the timeline is uncertain but said AI capabilities could advance far faster than expected.
“There’s a possibility we could be at the stage where these things are at the scary level of capabilities within six months,” he said. “That’s a real possibility.”
“One of the things that people are the most worried about is maybe it takes some sort of tragedy for people to take the dangers of AI seriously,” Coxon said.
You can watch Coxon’s full interview in the clip at the top of the page.
The post Ex-Anthropic researcher: Without AI guardrails, it could be ‘lights out’ for humanity appeared first on MS NOW.
