Andrew Ng on AI Extinction Warnings: Science Fiction or Reality?
September 17, 2026
Andrew Ng, co-founder of Google Brain and Coursera, has recently stated in an interview with Bloomberg TV that concerns about AI causing human extinction are largely science fiction. Ng argues that the industry often exaggerates these risks for publicity and to influence regulatory policies.
"Much more science fiction than science," Ng said, addressing the warnings of existential AI risk. He believes that the current hype around catastrophe is similar to what happened earlier in the AI boom, when an Anthropic researcher resigned, accusing his former company of risking people's lives.
In a US Senate forum in 2023, Ng had already expressed skepticism about AI extinction odds, putting them at roughly one in 10 million over a century. He claimed that large companies use fear to create regulations that hinder new entrants. Interestingly, Europe has already addressed these concerns in its AI Act, listing both the engineering risks Ng prioritizes and the loss-of-control risks he dismisses within the same legal framework.
The EU AI Act's list of systemic risks includes chemical, biological, radiological, and nuclear weaponization, cyber capabilities, disinformation, and major accidents in critical sectors. It also covers loss of control, aligning with Ng's concerns but also including models creating self-copies, which he considers fiction.
The obligations outlined in the AI Act are already in effect, with Henna Virkkunen, the EU's tech chief, emphasizing that Anthropic and OpenAI must assess the risk of losing control over their models, a duty that currently only applies above specific compute thresholds, affecting primarily large developers.