Anthropic Researcher Warns of AI Risks and Calls for Regulation

The Anthropic researcher Jacob Coxon, who is departing the company due to concerns that AI technology endangers human life, relayed to Fox News that there is optimism for avoiding a scenario where AI could harm humanity within the next decade. Discussing the issue on “Special Report,” Coxon, 27, emphasized the importance of regulation.

Coxon, in his resignation post on social media platform X, expressed that tech companies are not acting conscientiously and the race towards self-improving AI is reckless. With three years of pretraining research experience at OpenAI and Anthropic, Coxon believes the situation isn’t hopeless.

He told Fox News host Bret Baier that the threat is solvable. Coxon is confident that action can be taken to mitigate the risk of AI becoming a catastrophic threat. “It’s not like we think this is some impossible thing,” he stated, emphasizing the need for careful and regulated progression in AI development.

Anthropic’s alignment science lead, Evan Hubinger, backed Coxon’s concerns on X. “Jacob is correct here — we really do earnestly believe AI could kill all humans,” Hubinger commented, estimating a greater than 10% chance of this occurring within the next decade.

“To be clear, as we say in our latest Risk Report, I think the risk from present models is low,” Hubinger clarified. His concern focuses on superintelligence emerging from recursive self-improvement, which seems to be advancing rapidly.

Coxon concurs with Hubinger, stating the current threat is minimal but the danger lies in AIs improving themselves, possibly even next year. “This attitude is shared across the industry,” Coxon noted, as many realize the potential for AI dominance if development continues unchecked.

He called for international oversight, comparing the technological race to an armament competition. “This is possibly the most dangerous technology humanity has created,” claimed Coxon, advocating for international cooperation to prevent disaster.

Despite the risks, the potential benefits of AI are significant. Coxon believes we’re close to spectacular discoveries, stressing the need for a balanced approach. The industry needs regulation to prevent a hasty pursuit of progress, according to Coxon.

When asked by Baier whether the problem has a solution, Coxon responded affirmatively. “If we actually try instead of surrendering to the inevitability of a doomed race or a reckless gamble, then we can solve it,” he asserted.

Fox News Digital reached out to Anthropic for further comments.