Artificial intelligence researcher Jacob Coxon has resigned from Anthropic, raising concerns about the pace at which increasingly powerful AI systems are being developed and the safeguards surrounding them.
Coxon announced his departure in a post on X, saying he had spent approximately three years conducting pre-training research at both OpenAI and Anthropic.
In his statement, Coxon accused leading AI companies of moving towards increasingly capable, potentially self-improving systems without adequately addressing the risks they could create.
“Neither company is acting responsibly,” Coxon said, arguing that the industry was “gambling with our lives.”
Researcher warns about increasingly powerful AI
Coxon warned against underestimating how capable future AI systems could become.
He argued that sufficiently advanced systems could potentially outperform humans across important fields, discover cybersecurity vulnerabilities and gain access to significant resources.
His most striking claim concerned how some people working on advanced AI reportedly view the technology’s potential risks.
According to Coxon, people involved in developing AI genuinely consider catastrophic outcomes possible, including scenarios in which advanced AI could threaten human survival before the end of the decade.
His comments should, however, be understood as his assessment of AI risk rather than a prediction that humanity will be wiped out by AI.
Another Anthropic researcher backs concerns
Anthropic researcher Evan Hubinger subsequently expressed support for the broader concern.
Hubinger said he personally believes there is a greater than 10 per cent probability of AI causing human extinction within the next decade.
He also said that while he believes Anthropic is making efforts to address AI safety, researchers do not yet have a proven solution to the challenge of keeping hypothetical superintelligent systems reliably aligned with human goals.
The remarks add to an increasingly prominent debate within the technology industry over whether advances in artificial intelligence are moving faster than researchers’ ability to understand and control their potential consequences.
AI safety debate intensifies
Leading AI companies are continuing to develop systems capable of performing increasingly complex tasks, while governments, researchers and technology executives debate what safeguards should accompany those advances.
Concerns surrounding AI range from more immediate issues such as misinformation, cybercrime, job displacement and misuse to longer-term scenarios involving systems becoming difficult for humans to control.
Coxon’s resignation places him among researchers publicly calling for greater attention to those risks as competition among major AI laboratories intensifies.
Whether the most catastrophic scenarios ever materialise remains deeply uncertain.
What do you feel about this post?
Like
Love
Happy
Haha
Sad
