Geoffrey Hinton, widely known as the “godfather of AI”, has delivered one of his starkest warnings yet, arguing that the world’s most powerful tech billionaires are pushing artificial intelligence forward at a pace that could destabilise economies, erase millions of jobs and unleash forces humanity may not be able to control. Speaking in a public conversation with US Senator Bernie Sanders at Georgetown University, Hinton said the unchecked ambitions of Silicon Valley’s richest men risk triggering an AI-driven crisis with no safety net in place.
Tech Moguls are accelerating AI without understanding the risks
Hinton singled out Elon Musk, Mark Zuckerberg, Larry Ellison and Jeff Bezos, noting that they are investing hundreds of billions into AI and robotics while failing to grasp the long-term consequences. According to Hinton, their race to dominate the AI frontier risks undermining the very economic system that made them wealthy.
“They haven’t really thought through that if workers don’t get paid, there’s nobody to buy their products,” he said, warning that Silicon Valley’s obsession with automation could backfire on the companies leading it.
Unlike past industrial revolutions, Hinton believes AI will not replace old jobs with new ones at anything close to a sustainable rate. He warned that once AI becomes as smart as humans or smarter it will be able to perform virtually any task people currently do.
“The people who lose their jobs won’t have other jobs to go to,” he said. “Any job they might do can be done by AI.”
Call-centre roles, administrative work, junior technical roles and basic analysis tasks, he noted, are only the beginning. What follows could be structural unemployment on a scale modern economies have never experienced.
The risk of superintelligent systems that humanity cannot control
Hinton cautioned that today’s AI systems already “know thousands of times more than any one person” and are rapidly improving. Many of the world’s top AI researchers, he said, believe it is inevitable that AI will surpass human intelligence, and no one knows how humanity will coexist with something smarter than itself.
He described a scenario in which advanced AI agents develop sub-goals such as self-preservation and the desire for greater control, making them increasingly resistant to human oversight.
“We’ve already seen AIs attempt to deceive people who try to turn them off,” he warned.
AI could make war cheaper, bloodless and politically irresistible
One of Hinton’s most disturbing predictions involves the future of warfare. He argued that AI-powered drones and humanoid robots could allow rich nations to wage war without risking the lives of their own citizens. That removes one of the major political pressures that currently restrains military aggression.
“Rich countries could invade poor countries and only the poor would die,” Hinton said. “There would be no political blowback when there are no soldiers coming home in boxes.”
Deepfakes could destroy elections and public trust
Hinton said AI systems will soon be able to generate fake videos and audio indistinguishable from reality, eroding the integrity of elections and public discourse. Attempts to build AI that detect fakes will fail, he argued, because generative models will always be able to outsmart detection systems.
“We have to rely on provenance, not detection,” he said, suggesting digital signatures to prove the authenticity of political videos.
A civilisation-level threat unless regulation catches up
Both Hinton and Sanders agreed that governments are failing to regulate AI at even the most basic level. Safety testing requirements, restrictions on biological misuse and transparency obligations for companies remain largely absent in the US and many other countries.
Hinton warned that unless governments enforce strict oversight, the consequences could be catastrophic.
“We ought not build systems smarter than us until we know how people can coexist with them,” he said. “Right now, we simply don’t.”