AI staff 'genuinely frightened' for humanity's future, ex-Anthropic researcher tells BBC

What the report says

A former Anthropic researcher, Jacob Coxon, told the BBC that some people working on artificial intelligence were “genuinely frightened” about how quickly the technology is advancing and what that could mean for humanity. Coxon, who previously worked at OpenAI before joining Anthropic, said he resigned after posting online about AI risks and later speaking publicly about his concerns. He argued that the pace of development should be slowed and said the industry feels caught in a competition that makes staff uneasy about the consequences.

The BBC reports that Coxon’s comments came as Anthropic chief executive Dario Amodei also urged a slowdown in AI development, saying the risks are serious and that companies and governments need time to respond. Leaders at rival firms OpenAI and xAI have said they agree with some form of coordinated deceleration and regulation. Anthropic said it supports strong safeguards and a lawful, verifiable way for companies to work together on releasing powerful models.

Coxon said he believes an AI emergency could arrive within a short time frame, though such forecasts remain highly uncertain. He described fears among researchers about scenarios such as automated systems overwhelming parts of the internet. His remarks add to a broader debate in Silicon Valley and beyond, where some researchers and executives warn of existential risks while others question whether those warnings are overstated or tied to business and regulatory strategy.