'Godfather of AI' reveals frightening odds that AI will eventually 'seize control' from humans
Geoffrey Hinton, known by many as the 'Godfather of AI', has issued another warning surrounding the rise of artificial intelligence, indicating the frightening odds that technology will soon 'seize control' from humans.
Nobel Prize winner Hinton was one of the leading figures in the development of artificial intelligence, having previously led Google's DeepMind AI lab, but he has recently expressed his worries about how quickly the technology is developing.
One of the major fears Hinton expresses surrounding the widening wealth gap that AI will help create, as that could lay what he calls a 'fertile ground for fascism' as the rich increase their fortunes inversely with everyone else.
Advert
However, what he appears to truly find the scariest aspect of artificial intelligence is the prospect of its intelligence, illustrating that it poses a real threat of ending humanity in the near future.
In a new interview with CBS News, as reported by Business Insider, Hinton has indicated that AI could eventually manipulate and 'seize control' from humans as it increases its capacity for intelligence.
"Things more intelligent than you are going to be able to manipulate you," outlines the Godfather of AI, and with projections that we're less than a decade away from reaching artificial general intelligence (AGI), actions need to be taken to avoid the imminent threat.
He pessimistically joked that he's "kind of glad" to be 77 years of age, as that means he's likely to miss the worst of what AI will bring to the world, and compared the development of AI to raising a baby tiger.
"It's just such a cute tiger cub," Hinton began, but "now, unless you can be very sure that it's not gonna want to kill you when it's grown up, you should worry."
One Jeff Bezos-backed CEO has expressed similar worries about what AI could be capable of once it reaches or surpasses human intelligence, highlighting in particular the dangers of AI conducting R&D autonomously.
Yet Hinton has gone out of his his way to provide exact numbers that project the risk of an AI uprising, estimating a "sort of 10 to 20% chance" that AI could seize control from humans, but he does follow this up by arguing that it's hard to concretely predict.
Additionally, he points towards the dangers of AI agents, which perform tasks autonomously as opposed to the prompt-driven AI that's popular on a consumer level right now.
Meta CEO Mark Zuckerberg only recently revealed plans for AI to take over half of his company's coding tasks, where employees are moved to roles where they'll be 'managing' a team of AI agents, and this could very well be the start of what Hinton predicts to be a dangerous and potentially devastating future.