Former U.S. AI science envoy Rumman Chowdhury says research has not shown the extreme capabilities behind AI-extinction warnings, which she argues distract from everyday risks such as job loss and misinformation. She says public pushback gives her hope.
Her essay responds to a post by Anthropic researcher Jacob Coxon claiming AI builders believe the technology could kill everyone by the end of the decade.
Chowdhury points to Texas voters threatening to vote Democratic if Republican incumbents support data centers, and New York City's moratorium on AI in schools after parent groups and advocates objected to city guidelines they considered insufficient. She also cites polls showing three-quarters of people do not trust AI systems' outputs to be fair or accurate.
