In Pakistan, Anthropic researcher Jacob Coxon says he resigned over AI risks
In Pakistan, Anthropic researcher Jacob Coxon said he resigned, explaining that he was concerned about the race among artificial intelligence laboratories to create systems that, in his view, could pose a threat to human life within a few years. Dawn reports.
According to the publication, The Wall Street Journal was the first to report Coxon's departure. The newspaper described the case as one of the first examples of an Anthropic employee resigning over concerns about AI safety.
Criticism of AI developers
In posts on the social network X, Coxon said that for the past three years he had been conducting research into model pre-training at OpenAI and Anthropic. He stated that both companies are acting irresponsibly because, in his assessment, they are competing to create self-improving superintelligence.
More current news is available on the UA.News Telegram channel Telegram.
The researcher urged people not to underestimate the capabilities of this technology. According to him, such systems may soon surpass humans, hack anything, rapidly transform various industries, and gain real power and resources. Coxon also claims that progress in these areas is not slowing down.
Call for caution
Coxon said that some AI developers and executives privately fear that the technology could destroy humanity by the end of the decade. In his view, not everyone at OpenAI has recognized the civilizational risks, while Anthropic understands them but seeks to outpace other developers.
He called on laboratory employees to reconsider the pace of development and not launch reinforcement learning for superintelligent systems without a thorough understanding of their nature. As Dawn notes, Coxon was also among more than a thousand AI researchers who signed a statement calling for global coordination among governments to slow the development of models capable of improving themselves, if necessary.
