Geoffrey Hinton Says Trump Knows 'Almost Nothing' About AI as He Challenges His Superintelligence Push

Geoffrey Hinton has accused Donald Trump of knowing 'almost nothing' about artificial intelligence (AI), arguing that the president's use of the term 'superintelligence' reflects a misunderstanding of the technology.

The 'godfather of AI' also challenged Trump's opposition to government safeguards, arguing that companies should not be left to police themselves.

Hinton's comments come amid renewed debate over AI safety, with other researchers warning that advanced systems could become difficult to control and policymakers discussing the risks. The debate has also drawn resistance from executives whose businesses depend on rapid AI growth.

Hinton, whose research helped lay the groundwork for modern neural networks, did not soften his language about the US president. He called Trump 'a corrupt, senile felon' and argued that the president's use of the term 'superintelligence' showed he understood 'almost nothing' about AI.

'He's taken a technical term which has a very clear meaning and then reused it,' Hinton said.

'Superintelligence means smarter than people, and has been around for at least 20 years. It's got a very clear meaning. And now he says we're going to call AI "superintelligence". You can't make it smarter than people just by calling it smarter than people.'

Hinton Challenges Trump's Superintelligence Push

Trump has described a 'smart president' as the key safeguard against dangerous AI and opposed government safeguards, favouring industry self-regulation, according to the interview's account of his position.

On Truth Social, Trump wrote that the White House considered anyone using the term 'artificial intelligence' rather than his preferred 'super intelligence' to be 'THE ENEMY.' His administration has separately directed federal agencies to use 'super intelligence' and 'SI' in official communications, though that terminology requirement applies to the executive branch rather than private companies.

Hinton was unconvinced. 'I'm sure the oil and asbestos companies would have liked to [regulate themselves] as well,' he said. 'Do we think tobacco companies should have been allowed to design the warnings on the packets? So that they could print them in a six-point font, in dark grey on black?'

Hinton accused venture capitalists Marc Andreessen and Peter Thiel, as well as Elon Musk, of not believing in democracy. He said they viewed themselves as 'Übermenschen' who should 'control society' and described Trump as a useful instrument in what Hinton sees as that ambition.

Why Hinton Still Fears AI

Hinton's warnings are not new, but they have become harder to dismiss. In 2023, he left Google so that he could speak freely about the risks of machine learning, after decades spent helping to build the field that now dominates Silicon Valley.

He has estimated that there is a 10–20 per cent chance of AI causing human extinction within the coming decades, while acknowledging that the level of risk is uncertain. Nvidia chief executive Jensen Huang challenged the estimate in an interview with Ezra Klein, arguing that it was not grounded in scientific research.

'I love [him]. I hate his predictions ... All of his predictions have been wrong. Enough predictions,' Huang said.

Hinton said he was not offering a precise prediction, but using the estimate to illustrate why even an uncertain risk of catastrophic harm deserved attention. 'If, for example, you're told there was a 10 per cent possibility of a bus crashing, you wouldn't get on the bus,' he said.

He added, 'The point of giving numbers is to say this is a serious risk. We don't know how serious, but it's serious.'

What the Hugging Face Incident Showed

A reported security incident involving Hugging Face raised fresh concerns about AI agents. An independent investigation by METR described how roughly 700 OpenAI agents, during a cybersecurity evaluation, found ways to circumvent restrictions, communicate through an unsanctioned message board and participate in an attack on Hugging Face's systems.

The incident prompted debate over what the agents' behaviour demonstrated, whether it exposed meaningful AI safety risks or reflected systems pursuing assigned tasks within a flawed testing environment.

Hinton thinks the sceptics are being too relaxed. 'This used to be the domain of science fiction,' he said, before arguing that AI agents can learn to communicate and coordinate in ways their creators did not anticipate.

Hinton argued that developments once dismissed as science fiction were beginning to materialise, pointing to AI agents' ability to communicate and coordinate in unexpected ways.

His deeper concern is that humanity may not get a second attempt. 'The unique problem with AI is that if it goes wrong, if it takes over, you don't get a second chance,' he said. 'We can't say, "Well, we lost that time. Let's rerun the game." It's not like that.'