Geoffrey Hinton has warned that society may have been better off never developing artificial intelligence, while arguing that the technology should be subject to independent safety checks before companies release new models.
Speaking to The Guardian, Hinton said he does not consider himself an AI doomer and believes the technology could lead to positive outcomes. But he also cautioned that its development could have serious consequences. “It might be that we would have been better off never developing it,” he said.
Hinton proposed an oversight body similar to the US Food and Drug Administration, which regulates medicines. Under his proposal, AI companies would have to provide independent authorities with strong statistical evidence that a model does more good than harm before it is released.
He also criticized prominent technology leaders, including Mark Zuckerberg, Elon Musk and Sam Altman, saying he believes they understand AI’s risks but prioritize money. The comments were Hinton’s view, rather than a claim supported by evidence in the interview.
Hinton has previously distinguished between near-term risks, such as people misusing AI, and longer-term concerns about systems becoming highly capable and no longer needing human involvement. In a separate CNN appearance, he said AI could pose a genuine threat to humanity if researchers do not find ways to make it safe.
Despite those warnings, Hinton said he believes people may still be able to manage the technology, provided they make a substantial effort to do so.
Sources: Tom's Guide.