ARTIFICIAL Intelligence could kill all humans within the next decade, a technology firm’s safety chief has said.
US giant Anthropic’s Evan Hubinger put the level of risk at ten per cent.
He said the danger from current models was low but admitted he was worried the technology will improve itself to the point where it poses a global threat.
Hubinger spoke out after one of the firm’s researchers, Brit Jacob Coxon, quit, claiming tech firms were “gambling with our lives”.
Coxon left Anthropic, maker of the Claude chatbot, after three years training models for rival OpenAI, which runs ChatGPT.
He accused both of acting irresponsibly by allowing models to upgrade their own intelligence without human input, so-called “self-improving AI”.
READ MORE ON AI
BOT BEATER
How to defend your job from AI as thousands face redundancy – and get 34% raise
MEGA MIND
AI will find cure for cancer in our lifetime ‘that humans cannot’, tech CEO says
Coxon wrote: “They are racing straight to self-improving superintelligence and gambling with our lives.
“These will soon be superhuman systems that can hack anything, revolutionise any field overnight, and acquire real power and resources.
“The people building AI earnestly believe it could kill us all by the end of the decade.
“This is not a marketing stunt.”
Most read in Tech
FRESH APPLE
Apple unveils first foldable iPhone Duo that opens like a BOOK & iPhone 18 Pro
YABBA DABBA DOOBIE
Cavemen from 25k years ago were ‘first drug users’ & got buzz off nuts
OUT OF JUICE
Delete THESE 22 dodgy apps to save your phone’s battery life
ZUCK OFF
Woman says she has PROOF Facebook is listening in on conversations
Anthropic, OpenAI and Meta have all disclosed breaches by their AI tools as the race to build the most powerful goes on.
The Cabinet Office said the UK’s AI Security Institute was collaborating with the major AI firms “to make models safer”.