AI could kill all humans: Former Anthropic researcher warns of superintelligence risks

PTC Web Desk:  Artificial Intelligence could pose a bigger threat to humanity than nuclear weapons, according to former Anthropic researcher Jacob Coxon, who has warned that advanced AI systems could eventually become impossible for humans to control.

Speaking to Fox News, Coxon said the world already knows how nuclear weapons are controlled, but there is still no clear solution for controlling highly advanced AI.

“We don't yet know how to control AI,” Coxon said, describing it as potentially the most dangerous technology ever created by humans.

Coxon said he spent around three years researching AI at Anthropic and OpenAI. He believes the development of artificial superintelligence needs international cooperation and warned that an AI arms race could have serious consequences.

"We know how to control nuclear weapons... We don't yet know how to control AI."

Former Anthropic researcher Jacob Coxon, who recently left the company, is sounding the alarm on artificial intelligence, calling it "possibly the most dangerous technology that humanity has ever created."

Coxon, who says he spent three years conducting research at Anthropic and OpenAI, wrote in a viral post on X that the companies are more focused on beating each other to build the most advanced AI models than they are on safety.

He says an international AI arms race would be "disastrous," arguing global cooperation is the only way to manage the rapidly advancing technology.

@SpecialReport

— Fox News (@FoxNews) September 9, 2026

Warning over AI race

Coxon recently raised concerns on social media about the competition between major AI companies. He claimed that companies are under increasing pressure to develop more powerful models and are taking major risks in the process.

He said researchers working on advanced AI genuinely believe that future systems could pose an extreme threat to humanity.

Artificial superintelligence refers to a theoretical stage where AI becomes more capable than humans across a wide range of tasks.

Anthropic executive backs warning

Anthropic safety executive Evan Hubinger also acknowledged the seriousness of the risk.

In a post on X, Hubinger said the company genuinely believes there is a possibility that advanced AI could eventually cause the extinction of humanity. He estimated the risk at more than 10% over the next decade.

Hubinger said Anthropic is working to reduce the danger but admitted that there is currently no guaranteed way to ensure that an AI system more capable than humans would always follow the intentions of its creators.

He also described the risk from today's AI models as low.

Calls for govt action

Coxon said he believes companies cannot safely develop AI systems that surpass  human intelligence without stronger 
government involvement and a coordinated pause in development.

He acknowledged Anthropic's safety efforts but argued that competition between AI companies could push them to continue developing increasingly powerful systems even when the risks remain unclear.

Anthropic earlier removed a commitment from its safety charter that called for stopping model development if the company could not adequately control the risks. The company argued that stopping on its own could allow less cautious competitors to move ahead, potentially making the overall situation more dangerous.

The debate has also reached Washington. More than 1,000 technology workers, including Anthropic CEO Dario Amodei, have called for coordinated action to slow the development of the most advanced AI systems.

OpenAI also paused training on some of its latest models for two weeks in August before restarting under tighter safeguards. Its chief scientist, Jakub Pachocki, has separately called for extreme caution and urged governments to make international coordination on advanced AI a priority.

Also Read | 1980s AI photo trend: Bellbottoms, big hair & AI: Why your Instagram feels like 1985 right now