Anthropic Researcher Quits, Says AI Could Kill Us All

Jacob Coxon resigned from Anthropic over AI extinction fears, and a company team lead publicly confirmed believing the risk exceeds 10% within a decade.

@sharedot

Anthropic Researcher Quits, Says AI Could Kill Us All

By @sharedot · · 6 pages

  • AI
  • AI Safety

Jacob Coxon resigned from Anthropic over AI extinction fears, and a company team lead publicly confirmed believing the risk exceeds 10% within a decade.

A researcher walks out and says the quiet part aloud

Jacob Coxon announced his resignation from Anthropic in a post on X, writing that after three years of pretraining research at both OpenAI and Anthropic, 'neither company is acting responsibly' and that they are 'racing straight to self-improving superintelligence and gambling with our lives.'

According to Mashable, Coxon claimed the technology could kill countless people within the next ten years — if not everyone — and urged AI researchers to examine the ethics of their work and advocate for change.

His departure is the latest in a string of ethics-driven exits, with Anthropic safety lead Mrinank Sharma resigning in February and OpenAI researcher Zoë Hitzig quitting the same month over the company's advertising strategy.

The shock: Anthropic insiders agree the risk is real

What makes this resignation land differently is that Anthropic leadership did not deny it. active team lead at a frontier lab publicly endorsed an extinction-risk figure from a departing colleague is remarkable, and gives Coxon's warnings concrete institutional weight rather than leaving them as one researcher's opinion.

A pattern of departures behind the race narrative

Coxon is far from the first to leave a frontier lab over safety.

Per Mashable, Anthropic's safety lead Mrinank Sharma resigned in February warning the world is 'in peril' from AI and bioweapons, and wrote that he had 'repeatedly seen how hard it is to truly let our values govern our actions' inside the organization.

Also in February, OpenAI's Zoë Hitzig quit, publishing a New York Times op-ed raising 'deep reservations' about advertising built on ChatGPT users' most private conversations.

Coxon offered his own explanation for why labs keep building despite these fears: at OpenAI, many staff have not internalized the civilizational stakes, while at Anthropic the stakes are well understood but the company is locked in a race — believing no one else will act responsibly, so it must build first despite the risk.

The stakes for young people and what comes next

The warnings are already filtering into wider public concern.

KETV reports that Gen Z groups are sounding alarms about AI's impact on young people, with NoSo Collection Collective founder Maddie Freeman calling headlines about extinction threats 'horrifying' and noting that 67% of kids say they have used an AI chatbot.

Freeman said AI is complicating entry-level job hunting — with AI systems sorting hundreds of applicants per posting — and that some students are outsourcing their thinking entirely, creating a 'loss of creativity and agency.'

The backdrop is a competitive arms race: Mashable notes Anthropic CEO Dario Amodei has said superhuman AI could arrive by 2027, while OpenAI's Sam Altman expects AGI by the end of the year.

Coxon's call is for researchers to push for change from within; whether that pressure can slow a race this fast is the open question.

Sources

  1. mashable.com › Anthropic researcher quits, calls AI an existential threat to humanity
  2. ketv.com › US youth group warns of AI harms to kids and jobs amid Anthropic researcher quitting

More on AI