CHICAGO — Every invention of humankind — from the wheel to the forging of steel, from gunpowder to the automobile, to aviation, physics and the internet — has become weaponized. While each of these thresholds has advanced society, they have also brought us pestilence, genocide and war.
Geoffrey Hinton, the 2024 Nobel Laureate for physics, had three flights from Toronto to Chicago canceled Monday due to tornadoes, so he spent $2,000 on an overnight taxi ride to address the National Conference of State Legislatures.
Like the Blues Brothers, he was on a mission. His message about artificial intelligence, or AI, was sobering. No, it was chilling. Or as he put it, “scary” and even “terrible.”
“There’s a whole bunch of risks to deal with on people misusing AI,” said Hinton, of the University of Toronto. “In tech companies, it’s using AI to replace all the workers. The government will lose its tax base. It will be a disaster.”
This week, Bloomberg reported that major U.S. and global banks cut more than 10,000 jobs in the second quarter of 2026, continuing a broad workforce reduction trend driven by efficiencies, cost-cutting and the adaptation to AI. Reuters reported that Visa announced a 7% workforce reduction.
“We don’t know for sure what’s going to happen,” said Hinton, who in 2012 created the intellectual platform that has transformed AI into a ... being. Not a human being, but a creation of mankind that has transformed itself into a force now hellbent on not only survival, but the accumulation of power.
“Anyone who tells you they know what it’s going to look like in 10 years’ time is crazy,” Hinton said.
This comes as President Donald J. Trump has attempted to prevent the 50 states from regulating artificial intelligence. Trump’s 2024 restoration to power came with millions of campaign dollars from the so-called “tech bros” like Elon Musk, Peter Thiel and Jeff Bezos.
To Utah state Rep. Paul Cutler, “AI is here,” Cutler told the 1,000 legislators and 6,500 staffers and lobbyists attending the NCSL. “These technologies are already shaping work and science and society. As Congress has languished, state legislatures are stepping up. We’ve made it clear to Washington, states have a central role ensuring the safety of these technologies,” Cutler explained. “We have earned our voice at the table. States are neither pro- nor anti-AI. We are looking to capture the promises of these technologies while minimizing the risks.”
Despite Hinton’s wry quips that brought temporary comic relief, it may all be coming too late.
“This is totally new territory,” Hinton explained. “We’ve already seen lots of bad things happening,” like getting politically divisive echo chambers; Facebook and YouTube showing reels “more extreme because it keeps your attention”; the potential for corrupted elections; the probability of mass surveillance on the population; and increasingly more sophisticated cyberattacks.
“They can very easily make deadly viruses,” Hinton said, which could be aimed at specific ethnic groups or genders. “That’s very scary.”
“And then there’s this new risk altogether, which is the risk these new beings will actually take over,” Hinton said, repeatedly referring to “AIs” as “they.”
“Nobody knows how to prevent that,” he said. “We have every reason for thinking they may try to take over. All you do is give it the ability to create itself; as soon as it has the ability to create self-goals.”
“If you want to make an effective AI agent, you will give it a goal, and it figures out how to make its own goals,” Hinton explained. “Once it has the ability to make its own goals, the first one is to gain more power. If you gain more power or control, you’re better at achieving your other goals. They just want to achieve more power.
“Another obvious goal is to survive.” Hinton continued. “We know they’re very good at being patient. They will be able to manipulate people they talk to, for example, into not turning them off. They will be very convincing.
“So that’s the end.”
The sprawling McCormick Place hall exploded in laughter. But it was the kind of emotional eruption that had legislators shaking their heads and wincing.
Cutler asked Hinton how AI might improve our lives.
“It’s got huge benefits,” Hinton explained, citing “huge gains in productivity” and the potential advances in healthcare and education. On the medical front, AI is already making diagnoses better than human beings. He said in the United States and Canada, 200,000 people annually die from bad medical diagnoses. “AI should be able to at least halve that.”
Cutler conjured the notion in the not-too-distant future of “loved ones not witnessing a relative dying of cancer or heart disease.”
Is it too late to prevent the evil AI?
Hinton dropped this stunning scenario: “Suppose there’s a big-time terrorist and you hear he’s in a school. Are you willing to drop a bomb on the school? Some people say you should never do that, but from my point of view, it’s 1938, and if I knew what was going to happen next, and Hitler was in the school, I might be the one to destroy the school to get rid of Hitler.”
That is an Odesseyian dilemma: sacrificing 1,000 kids and teachers to avoid a world war that killed 60 million ... people.
“It would be a difficult decision, right?” Hinton asked. “But it’s not clear you shouldn’t do that. So this is an example where people don’t align.”
The epochal question facing legislators from 50 states is how human beings will interface with “them” — the AI we have created but appears to be leaping out of the sandbox.
Brian A. Howey is an opinion columnist for State Affairs Indiana and the founder of Howey Politics Indiana. His writing offers analysis and opinion shaped by decades of experience covering Indiana politics. Email him at howey@stateaffairs.com.
