Nvidia Chief Rejects AI Extinction Forecast as Safety Debate Sharpens
Nvidia co-founder and chief executive Jensen Huang says artificial intelligence will not drive humanity to extinction by 2030.
In an interview with CBS News reported on Sept. 21, 2026, Huang said: "2030 is not going to be the end of the world. There is 0% chance that's going to be the end of the world" The Guardian. He rejected extinction warnings as overblown 'doomsday narratives' and said they are 'not grounded in science'. CBS News reported that Huang disagrees with recent warnings from AI researchers that the technology could lead to humanity's extinction CBS News.
His comments were directed at former Anthropic researcher Jacob Coxon, who said on social media that AI could become superhuman — able to outperform people across many tasks — and kill off humanity within the decade. Coxon's remarks were backed by two other researchers at Anthropic, the maker of the AI assistant Claude. Huang called the claim 'irresponsible'. He said scaring people about AI is unnecessary and irresponsible.
Huang also said: "However it's characterised, 2030 is not going to be the end of the world", according to Sept. 19 reporting by The Straits Times The Straits Times. He stressed safe management of AI while dismissing doomsday fears. On pace, he said: "We should go as fast as we can, irrespective of anyone else" and rejected new AI regulations, meaning binding rules on how systems are built and released Tom's Hardware. He dismissed Anthropic's warnings about AI-caused doom. The New York Post reported on Sept. 20, 2026 that Huang assured there is a '0% chance' AI will end the world by 2030.
Anthropic CEO Dario Amodei reiterated his call for a slowdown of AI development at a conference in San Francisco. The Guardian published the report 'Anthropic CEO renews call for AI slowdown as Nvidia's urges ...' on September 15, 2026. Amodei, OpenAI boss Sam Altman, and Elon Musk have called for development of AI technology to be slowed down. US lawmakers criticised AI companies over AI risks.
Anthropic published a threat intelligence report, essentially a log of misuse attempts it observed. It detailed attempts by criminals, state-sponsored groups, spyware vendors, scientists and propagandists to use its AI models to design missiles and bombs, create deadly pathogens and spy on dissidents.
Donald Trump dismissed anxiety over AI development as a 'hoax' and a 'conspiracy' and said a slowdown could give China the edge. US Treasury Secretary Scott Bessent said the US and China had agreed to set up an AI dialogue mechanism, a channel to alert each other, to help contain AI's potential dangers.
The broader context here is a clash between two risk vocabularies. One treats extinction as a planning assumption that warrants restraint and enforceable rules. The other treats it as a distraction from deployable risks that can be managed through engineering controls and bilateral notification. Both sides cite the same misuse evidence and draw opposite procedural conclusions.
Looking at what this means for diplomacy, the operative question is not who wins the rhetoric. It is whether a US-China alert mechanism can function while lawmakers debate constraints, Trump warns against ceding advantage to Beijing, labs document misuse they say requires slowing down, and Huang argues for maximum speed. That configuration leaves verification, thresholds and incident-sharing undefined, and puts private threat reporting at the center of public policy without agreement on what triggers action.
