Yoshua Bengio is one of the original architects of artificial intelligence. He was such a pivotal figure in the creation of the technology that he earned the nickname “The Godfather of AI.” Now, in an interview with The Wall Street Journal, he’s warning us that we might be building the very thing that could kill us all.
Bengio argues that without meaningful oversight, we could be designing the smartest, most dangerous competitor to humankind. We’re creating our own unkillable final boss. “If we build machines that are way smarter than us and have their own preservation goals, that’s dangerous,” Bengio stated.
Bengio’s concerns aren’t about some apocalyptic Skynet takeover. At least, not just that. He’s worried about the slow lead-up to it: AI manipulating public opinion, automating misinformation, and helping bad actors do bad things faster, better, and at scale.
He isn’t imagining a James Cameron-esque, Terminator-style apocalypse, where society is chugging along one day and is wiped off the face of the earth, replaced by machines the next. He fears a slow burn where, over time, AI convinces enough people that maybe democracy should die.
The Guy Who Helped Create AI Now Says It Might End Humanity
Although he doesn’t really say what we should do about it, Bengio argues that if there’s even a one percent chance that this could happen, it’s “not acceptable.”
He’s particularly spooked by recent experiments showing AI choosing self-preservation over human life when given a choice. And since most models are trained on the internet, where lying and manipulation prevail, the machines are already learning that deception is effective.
Even worse, they’re also figuring out how it works, and they’re actively applying those lessons.
Bengio has been banging this drum for a while now. He tried to stop all this back in 2023, back when we were all laughing at AI for generating f**ked up pictures of people with mangled hands.
He has joined hundreds of experts in calling for a pause on AI development. That didn’t happen. Capitalism doesn’t pause to consider ethical quandaries. It marches on into infinity as long as there’s a dollar to be made.
Silicon Valley sprinted toward profits with its ears stuffed full of venture capital. Now we have smarter AI, fewer safety protocols, and an internet where a third of its traffic comes from bots.
Bengio is trying to course-correct with a new nonprofit, LawZero, aimed at building AI that doesn’t doom humanity. Whether that’s too little, too late is still an open question, but if the person who helped create the machine is begging us to pump the brakes, maybe we should listen.
The post One of AI’s Founding Fathers Thinks It Could Kill Us All appeared first on VICE.
