News presenter left lost for words as 'Godfather of AI' confirms says AI could wipe out humanity
Recent fears surrounding the potential for AI to end humanity in the coming years have left one news presenter stunned, with Geoffrey Hinton – colloquially referred to as the 'Godfather of AI' – confirming the bad news in a new interview.
It all began when ex-Anthropic and OpenAI researcher Jacob Coxon dropped a bombshell on social media this past Tuesday, as he revealed how concerns over the future of civilization led him to resign from the former company.
While his concerns over the ethical and moral considerations at both AI development labs have rightfully drawn plenty of headlines, the biggest conclusion people have taken from Coxon's declaration is that superhuman AI holds the potential to 'kill us all' by the end of the decade.
This isn't a worry that originates from Coxon alone, however, with Hinton being among the leading figures to repeatedly warn about the dangerous potential of advanced AI technology over the years, and he has reaffirmed these beliefs to the shock of a news presenter.
News presenter left shocked by Hinton's AI claim
Hinton took part in an interview on BBC's Newsnight where he discussed the claims made by Coxon, and presenter Victoria Derbyshire asked him specifically about the threat of significant loss of life.
Advert
She highlighted a particular prediction that suggested there's currently a 10 percent chance of human extinction, and while Hinton advised caution when ascribing a specific figure to something that is completely new to us, he concurrently indicated that "a ten percent chance seems not an unreasonable estimate to me."
Derbyshire was understandably taken aback by this claim, asking the highly-respected figure to reaffirm whether this was what he was claiming — yet once he confirmed the risk again she was left stunned.
"Wow. Oh my god," she uttered, likely reeling in fear for the dangers that currently face global civilization thanks to these rapidly advancing tools.
How could AI end up killing humans?
Beyond being aware of the alarmingly high risk of global extinction, it's important to also understand how and why the continued development of AI tools could lead to this point, and Hinton offered his own perspective.
"It would even have to act 'in the real world' to cause devastation," he explained, "it would be able to cause complete chaos just by talking to people. But it wold also be able to design very nasty viruses – biological viruses – or it could be computer viruses. It could do devastating cyber attacks."
This is something that OpenAI CEO Sam Altman has specifically warned about, suggesting that the latter could arrive as early as this year with new research suggesting that the latest versions of Claude are already technically capable of this.
"There are countless ways it could get rid of us if it wanted to. We have to figure out how to design it, so it won't want to."
The danger appears to specifically arrive when AI becomes fully autonomous, as humans will begin to lose control when it solely drives its own development and also decides what its goals are, giving us little choice if it goes off the rails.