‘Very probably’ AI might quickly wipe out a whole inhabitants, claims tech boffin

One of the world’s leading AI experts has said robots could soon launch ‘a million weapons at once’ without needing human supervision, which could be catastrophic

Terminator’s Judgement Day is looming for planet Earth – with a Brit tech boffin warning it is “very likely” artificial intelligence will “soon wipe out an entire population”.

There have already been some slightly less terrifying warnings about AI. Some believe AI will end online dating, while others have warned that “a red line has been crossed” as it “can now replicate itself”.

Now, computer scientist Stuart Russell, one of the world’s leading artificial intelligence experts, has said bots could launch “a million weapons at once” without human supervision or conscience. The result could be a catastrophic destruction of mankind straight out of The Terminator movie series in which cyborgs take over the Earth.

Stuart, who heads the Center for Human-Compatible Artificial Intelligence at the University of Berkeley California, US, is calling on governments worldwide to unite to ban ‘bot weapons.

He said the “autonomous” killers were already operating in the Russia-Ukraine conflict.

“The main concern is that because AI enables a weapon to operate entirely without human supervision that means that one person could push a button and launch perhaps a million weapons at once and wipe out an entire population,” he said

“We don’t think that is a good idea. But that seems to be the direction that things are going right now.”

Asked by BBC’s Newsnight what he believed were the chances of such destruction happening he replied: “To be honest I think it’s very likely.”

He said Ukraine had become a “crucible for technological advancement in these kinds of weapons” which were becoming “more and more autonomous” due to the development of “electronic counter measures”.

“It used to be that both sides were killing each other using remotely-piloted drones which require a radio link so that the operator can see through the camera,” he said. “But then with jamming those links are starting to become impossible.

“So if you want your weapon to do something useful you have to make it fully autonomous. That has already happened.”

Professor Russell, 63, from Portsmouth, Hants, who penned the world-renowned university textbook Artificial Intelligence: A Modern Approach, said many nations agreed ‘bot weaponry should be outlawed.

But he said some superpowers were against a ban.

“There are major powers, including the US and Russia, surprisingly collaborating on preventing any discussion about banning these kind of weapons,” he added.

“Most people right now when they think about AI think about good things like medical diagnoses, having someone to play chess with. And we, the AI community, don’t want the death of human beings to be the first thing that comes to mind when we think about AI.”

For the latest breaking news and stories from across the globe from the Daily Star, sign up for our newsletters .