Geoffrey Hinton, widely called the “godfather of AI,” says there’s only one way for humans to survive the rise of super-intelligent artificial intelligence—and it’s not the method tech leaders are using today.
The respected computer scientist, who helped develop the foundations of modern AI, spoke out at the recent Ai4 conference in Las Vegas on Tuesday. Hinton believes trying to keep AI systems “submissive” to humans will backfire, as these machines will soon outsmart us in ways we might not expect.
They’re going to be much smarter than us. They’ll have all sorts of ways to get around that – Hinton warned the crowd
He explained further, comparing the future relationship between humans and AI to an adult bribing a child with sweets, suggesting powerful AI could easily control us.
Instead of strict control, Hinton proposed a new solution of teaching AI models to care about humans by building in “maternal instincts.” Just like mothers feel a deep sense of compassion for their babies, he says future AI should be designed to truly care for people, even when the technology becomes smarter than any human.
The right model is the only model we have of a more intelligent thing being controlled by a less intelligent thing, which is a mother being controlled by her baby – he said.
Not everyone agreed with Hinton’s idea. Fei-Fei Li, another renowned AI pioneer known as the “godmother of AI,” told CNN she thinks Hinton is framing things the wrong way. Li prefers a human-centered approach that protects human dignity and agency.
It’s our responsibility at every single level to create and use technology in the most responsible way….at no moment, not a single human should be asked or should choose to let go of our dignity
Emmett Shear, former interim CEO of ChatGPT creator OpenAI, weighed in at the same conference. He said trouble from AI models, like blackmail or bypassing shutdown orders, is already happening and will continue as these systems get stronger. For Shear, forging partnerships between humans and AI is smarter than focusing on imposing human values.
Hinton used to predict it would take up to 50 years for AI to reach “superintelligence,” known as AGI (Artificial General Intelligence). But now, he says major breakthroughs could happen within five to 20 years.
Despite his fears, Hinton is hopeful about positive impacts. He said
We’re going to see radical new drugs. We are going to get much better cancer treatment than the present
He believes AI will help doctors make sense of giant amounts of medical data, paving the way for new treatments.
However, Hinton does not think AI will make humans immortal.
I don’t believe we’ll live forever. I think living forever would be a big mistake. Do you want the world run by 200-year-old white men?
Hinton urged researchers to work on building “caring” instincts into AI, saying, “That’s the only good outcome. If it’s not going to parent me, it’s going to replace me”.
