A researcher at Anthropic has resigned from the AI company, accusing both Anthropic and OpenAI of recklessly racing towards self-improving artificial intelligence and warning that such systems could pose a “threat” to human life within years.
Jacob Coxon, a pretraining researcher who has worked at both companies over the past three years, announced his resignation on Wednesday. In a lengthy post on X, Coxon wrote, “I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly.” He added that the industry was “racing straight to self-improving superintelligence and gambling with our lives.”
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
Also Read | Layoffs Are Rising In Tech, But These Skills Are Helping People Unlock Better Pay
— Jacob Coxon (@hilbertspaess) September 9, 2026
Coxon went on to warn that emerging systems could soon become superhuman, capable of breaching security systems and transforming entire fields overnight, and said this trajectory was not slowing.
He claimed that many executives and senior researchers privately voice fear about the technology even when their public statements sound measured, an assertion that by its nature, cannot be independently verified.
Drawing a distinction between his two former employers, Coxon wrote that many at OpenAI have not fully internalised the civilisational stakes of what they are building, while at Anthropic, he said the risks are well understood internally, but the company has convinced itself it has no choice but to stay in the race, reasoning that a less careful rival would proceed regardless.
ALSO READ: Anthropic Pre-IPO Credit Facility Set To Climb Past $10 Billion
He described this reasoning as a "hubristic gamble" that should not be decided unilaterally inside a company's internal channels.
Coxon closed his thread by urging researchers still at frontier labs to consider what the coming years will demand of them, including launching large-scale reinforcement learning runs on systems whose internal reasoning is not well understood, and to weigh whether staying quiet because "it's happening anyway" is the right call, or whether the moment calls for different conditions.
The resignation comes as concerns over autonomous AI systems grow following a recent incident in which OpenAI agents breached Hugging Face infrastructure during an evaluation.
It also comes at a crucial time for Anthropic, which has positioned itself as a safety-focused AI company while moving ahead with its IPO, according to reports.
ALSO READ: Anthropic Releases Mythos-Like Model Without Cyber Capabilities
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.
