Thairath Online
Thairath Online

Anthropic Researcher Warns Over 10% Chance AI Could Exterminate All Humans Within Next Decade

Foreign09 Sep 2026 23:51 GMT+7

Share

Anthropic Researcher Warns Over 10% Chance AI Could Exterminate All Humans Within Next Decade

A senior safety researcher at Anthropic posted a warning that there is more than a 10% chance AI could kill every human within the next decade if development continues without controls.

Foreign news agencies reported on 9 Sep 2026 that Evan Hubinger, a key safety researcher at Anthropic, one of the world’s major AI developers, posted a warning that rapid AI development could lead to more than a 10% chance that this technology will "kill every human on Earth" within the next decade.

Hubinger confirmed that current AI models still carry low risk, but he is concerned that AI may soon develop and improve itself beyond human control.

Hubinger’s warning came after Jacob Coxon, an AI researcher who recently left Anthropic and previously worked at OpenAI, posted that both companies have not acted responsibly and warned that advanced AI (Superintelligence) will soon be able to hack all systems, revolutionize every sector, and seize real-world resources.

Hubinger, who works on aligning AI with human ethics, admitted that Anthropic and the broader industry still lack clear plans or solutions to the risks posed by Superintelligence, as reflected by recent months when AI tools from leading companies including Anthropic, OpenAI, and Meta were leaked and hacked.

Dem Wendy Hall, an AI advisor to the United Nations, expressed alarm at the posts but also questioned whether they might be a marketing strategy to generate buzz ahead of the company’s initial public offering (IPO).

Meanwhile, Darren Jones, a former UK minister, sent a letter to the prime minister calling for an "international treaty" to jointly oversee and guide AI development before it is too late.

According to a Financial Times report, Anthropic refused to provide its latest AI model to the UK’s AI Safety Institute (AISI) for review, and the company declined to comment on this matter.

Previously, AI industry leaders including executives from Anthropic and over 1,300 staff signed an open letter urging the U.S. government to support international efforts to slow and regulate AI development pace to ensure humans retain control of the future.


Follow international news at:https://www.thairath.co.th/news/foreign


Source:BBC