AI has a 10-20% chance of destroying humanity in the next three decades
Geoffrey Hinton was on the show's microphone Today from BBC Radio 4. According to him, AI has a 10 to 20% chance of destroying humanity in the next three decades. For what ? Because we have never had to manage entities more intelligent than us.
AI can destroy us according to Geoffrey Hinton
“How many examples do you know of where a less intelligent thing controls a more intelligent thing? There are very few examples. There is the mother and the baby. Evolution has worked a lot to allow the baby to control the mother, but that's the only example I know of”he explains.
The University of Toronto professor emeritus compared humanity facing AI to young children: “Imagine yourself facing a three-year-old child. We will be three-year-old children.” For him, the technology has exceeded his expectations: “I didn't think we'd be here now. I thought we'd get there at some point in the future.”
To express himself more freely on the risks of AI, Geoffrey Hinton did not hesitate to leave Google last year, worried about the rapid advancement of technology. “The current situation is that most experts in the field believe that in the next 20 years we will develop AIs that are smarter than humans. This is a very scary thought”he emphasizes. “The development is very, very fast, much faster than I expected.”
The scientist urgently calls for world governments to regulate this technology. According to him, the market is not able to ensure secure development despite the measures put in place by companies like OpenAI. “My concern is that the invisible hand will not keep us safe. Letting big business solely pursue profit motives will not be enough to ensure safe development.”he says. “Only government regulation can force these big companies to do more research on security.”
Yann LeCun, also described as “godfather of AI”has a completely different position. As Meta's chief AI scientist, the expert downplays the threat to humanity and believes that technology, on the contrary, “could actually save humanity from extinction”. Geoffrey Hinton is one of three “godfathers of AI” awarded the ACM AM Turing Prize, the equivalent of the Nobel in computer science.
