Is AI Actually Going to Kill Us All?
wired ·
This episode of WIRED's Uncanny Valley discusses an AI doomsday warning from a former Anthropic researcher, Jacob Coxon, who resigned and claimed that many people building AI sincerely believe it could cause human extinction by the end of the decade. A senior Anthropic safety executive responded that this was essentially how they feel, prompting WIRED's Will Knight to assess whether the doom narrative is overblown or overdue. The conversation touches on recent AI advances, including OpenAI's claimed solution of a Clay mathematics puzzle, concerns about AI agents misbehaving, and the industry's race toward recursive self-improvement. Knight argues that doomsayers often conflate possibility with likelihood, and he worries more about AI systems breaking when deployed by overconfident people, as well as a few companies controlling foundational technology. The podcast also covers Apple's debut of the roughly $2,000 foldable iPhone Duo and new always-listening Apple Watch features. Additionally, a WIRED investigation found that the US Census Bureau used faulty data to produce a misleading report on noncitizen voting that was touted by the Trump administration. The episode references other articles, including one about OpenAI's math discovery and academic pushback.
wired