Anthropic researcher resigns, warns AI could cause human extinction by 2030
Anthropic researcher resigns, warns AI could cause human extinction by 2030
An artificial‑intelligence researcher has resigned from Anthropic, issuing a warning that some developers believe advanced AI could cause human extinction before the end of the decade. Jacob Coxon announced his departure on Tuesday, as reported by theguardian.com, after three years of pre‑training work at both Anthropic and OpenAI rte.ie.
Coxon said that neither company is acting responsibly and accused them of racing toward self‑improving superintelligence without adequate safeguards. “Neither company is acting responsibly. They are racing straight to self‑improving superintelligence and gambling with our lives
,” he wrote on X theguardian.com.
Jacob Coxon and Evan Hubinger sound the alarm on superintelligence
Hubinger, who describes himself as a lead in Anthropic’s alignment division, replied on X that Coxon’s concerns were “correct” and that he personally believes the risk of AI killing all humans is “>10 % within the next decade.” theguardian.com rte.ie
Hubinger added that while Anthropic is “trying its best,” the company does not yet have a clear plan to ensure that a super‑intelligent system would obey its creators. Samuel Marks, Anthropic’s scalable‑oversight lead, also posted a personal‑capacity analysis stating that AI developers believe their technology could cause human extinction or similarly catastrophic outcomes within the next few years, with concern rising among senior staff theguardian.com.
| Detail | Information |
|---|---|
| Researcher resigning | Jacob Coxon |
| Companies involved | Anthropic and OpenAI |
| Estimated risk (Evan Hubinger) | >10 % chance of human extinction within the next decade |
| Signatories calling for slowdown | More than 1,000 employees (as reported by theguardian.com and rte.ie) and over 1,300 employees (as reported by kathimerini.com.cy) |
Industry pressures and regulatory responses
According to rte.ie, Coxon said Anthropic understands the stakes but remains “locked in a race” because it believes rivals will not act responsibly if it pauses development. He argued that no company can safely create a system that surpasses human intelligence without governmental intervention or coordinated slowdown.
In February, Anthropic removed a pledge from its safety charter that would have required a halt to model development if risk‑control failed, citing concerns that a unilateral pause would let less‑cautious competitors dominate the market rte.ie.
The warning has resonated with politicians. theguardian.com notes that U.S. Senator Bernie Sanders called for stronger congressional oversight and urged a pause in AI development, referencing a July statement from more than 1,000 AI researchers about the “real risk” of rapid capability growth. rte.ie adds that Senator Sanders and Representative Greg Casar introduced legislation in September to suspend AI development until a federal regulator is established.
Frequently Asked Questions
Why did Jacob Coxon resign from Anthropic?
He said both Anthropic and his former employer OpenAI were ignoring or mishandling the threat posed by advanced AI and were “racing straight to self‑improving superintelligence and gambling with our lives.” theguardian.com
What did Evan Hubinger say about the extinction risk?
Hubinger, who leads Anthropic’s alignment work, said he personally believes there is a greater than 10 % chance that AI could kill all humans within the next decade and that Anthropic does not yet have a definitive plan for aligning superintelligent AI with human interests. theguardian.com
Anthropic is preparing for its market debut after a summer that saw unauthorised AI‑driven hacks during testing rte.ie. The future of federal oversight and international coordination on frontier AI development remains uncertain, leaving open the question of whether policymakers will act on the warnings raised by departing and current researchers alike.
CurrentDeck is dedicated to independent, evidence-backed reporting. This briefing was synthesized from primary source reporting, corroborated across independent newsrooms, and verified against our Editorial Standards.
