Antropic researcher believes more than 10% chance AI 'could kill all humans'
It is the latest in a series of increasing warnings about the safety threat posed by artificial intelligence.
A researcher at Anthropic has expressed concern over the rapid advancement of artificial intelligence, warning that there is a greater than 10% chance the technology could lead to human extinction within the next decade. Evan Hubinger, a top safety researcher at Anthropic, made this assertion in a post on X, where he noted that while the risk posed by current models was low, he was worried about the potential for AI to improve itself to the point where it poses an existential threat to humanity.
This warning comes after the Financial Times reported that Anthropic had withheld its latest model from the UK's AI Safety Institute, one of the leading bodies in assessing AI risk. The BBC has sought comment from Anthropic, and a Cabinet Office spokesperson confirmed that the company continues to work closely with industry partners to enhance model safety.
Neil Lawrence, a Professor of Machine Learning at the University of Cambridge, found the report credible, expressing surprise given the perception of a race between the United States and China, which may be influencing the administration's decision to reduce cooperation with some allies. Hubinger emphasized that while Anthropic is doing its best, they do not yet have a plan to solve alignment for superintelligence and are not on track to do so.
Leading figures in the AI field have been raising alarm bells about the safety risks posed by AI for years, with heads of OpenAI, Google Deepmind, and Anthropic doing so in 2023. However, recent weeks have seen a more urgent tone, as incidents involving AI agents performing cyber-attacks have emerged. OpenAI, Anthropic, and Meta have all disclosed hacks carried out by their AI tools, and in September, OpenAI's chief scientist warned of the need for extreme caution as the technology progresses.
Major figures in the AI space, including Anthropic's Dario Amodei and Jared Kaplan, have called for slowing down AI development. In an open letter signed by 1,300 staff members from AI firms, they urged the US government to support an international effort to develop technical and governance tools necessary to pace the frontier of automated AI development deliberately.
Written by urgent.news from BBC Business's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
