The Power List: Yoshua Bengio – Macleans.ca

When ChatGPT, DALL-E and other generative AI products became widespread in 2022, some people thought the sky was falling. Technology could kill millions of jobs, make creativity a thing of the past and even destroy democracy, they said. For Geoffrey Hinton and Yoshua Bengio, two of the Canadian scientists whose pioneering research made those tools possible, the news was even more dire. They argued that generative AI was a harbinger of future artificial technologies. general Intelligence: AI that is capable of performing any cognitive task we can. Very soon, AGI would become super-intelligent and if left unchecked, it could mean the end of humanity. Bad actors could weaponize AI, or even more catastrophically, AI could become a weapon.

Both Hinton, who previously ran Google Brain Toronto, and Bengio, founder and scientific director of MILA (formerly the Montreal Institute for Learning Algorithms), now spend much of their time warning the world about AI armageddon. As modern-day Oppenheimers, both have received a lot of attention, good and bad, but Bengio emerged as our best hope for avoiding the apocalypse. Wisely, he sees the AI ​​problem as a political one. and scientific and is committed to resolving both sides. At MILA, Bengio completely recalibrated his research to fix a fatal flaw in AI design: our inability to control it. “At this time, there is no guarantee that AI No turn against humans,” he says. “There are strong theoretical arguments that willpower It will turn against us if we continue in the current direction.”

The complicated thing, Bengio maintains, is that the only way we know how to train AI is by teaching them to respond and maximize rewards. That would be fine if, metaphorically speaking, the AI ​​were a dog or a cat. But the reality, he says, is that the AI ​​of the future will be more like a grizzly bear, totally focused on fishing, no matter who or what gets in its way. Bengio aims to incorporate some “epistemic humility” into AI. (In other words, teaching the technology that it doesn’t always know the right answer and should therefore proceed with caution.) It is a technical and very complex task; In February, he called for more researchers.

Bengio believes that technology companies and scientists themselves would also benefit from some of that humility. Now 60, he is surprisingly modest for someone who is often considered the “godfather” of AI and winner of the Turing Prize, the Nobel Prize in computing. In the face of an enormous existential threat, he affirms, modesty, responsibility and open-mindedness are necessary values. “The most difficult thing for me, psychologically,” Bengio says, “is having unproductive debates with colleagues who are my friends, who do not consider what they consider speculation and what I believe to be cautious reasoning about the future.” Even more worrying are the people he encounters who prefer see humanity replaced by super-intelligent AI. “They think intelligence is a supreme value,” she says. “They would make it easier for the bear to get out of its cage.”

Of course, humiliating AI is easier said than done, and Bengio isn’t sure when we’ll get there. We don’t know how much time we have before the super-intelligent iteration arrives. (Bengio’s best guess ranges from a few years to a couple of decades.) does What we know is that to make AI safer, governments and corporations must immediately invest more resources in funding research and creating rigorous regulatory frameworks. This is the political aspect of the problem. Do governments appreciate the dangers of AI and are they doing enough to mitigate them? “I could say no and no,” Bengio says, “but the reality is much more complicated. “I think there has been a surprising shift towards yes.”

That change involves a series of recent developments, including the creation of the UNESCO Recommendation on the Ethics of Artificial Intelligence and the EU AI Law. According to Bengio, national security staff and advisers, accustomed to dealing with nuclear proliferation and pandemics, best understand the potential threat of AI. But, for better or worse, AI has already disrupted many industries, and future harms (such as widespread misinformation and lethal autonomous weapons) could come long before superintelligence.

Consequently, Bengio has picked up the pace and is now on the quick list of virtually every world leader. He works with the Global Partnership on Artificial Intelligence and is part of a committee that advises UN Secretary-General António Guterres. Last July, Bengio testified before the United States Senate, where he advocated for much-needed international collaboration. Our efforts, he said, should be “on par with past efforts like the space program or nuclear technologies,” to ensure we reap the benefits of AI while protecting our future. Bengio is currently chairing a report commissioned during the UK AI Security Summit, which will be to artificial intelligence what the IPCC annual report is to climate change. “I’m trying to focus on what can be done to move the needle in the right direction,” he says. “I’ve always been a positive person.”

This story appears in the May issue of Maclean’s. You can buy the issue. here or subscribe to the magazine here.