Geoffrey Hinton, the computer scientist known as AI's 'godfather', says the models he helped build are alien beings that genuinely understand what they're saying - and that humans now have only a couple of years to figure out what we owe them.
In a recent interview with The Guardian, Hinton argued that AI systems "understand what they're saying" because, in his view, "you couldn't possibly answer all the questions without understanding the question." He went further, calling today's models "beings of a different kind" that deserve consideration beyond raw intelligence, even while insisting humans should still come first in any "symbiotic relationship" with them. Hinton, who left Google in 2023 to speak more freely about AI risk, also took a swipe at the industry's self-regulation efforts, likening it to letting tobacco companies design their own warning labels.
Hinton's remarks land in the middle of a louder fight over how AI should be governed - Microsoft's president has complained that a carton of milk faces tighter rules than AI systems do. Philosopher Jacob Fox offered a more skeptical read: he argued Hinton gave no good reason to believe AI has subjective experience, and that framing personhood around behavior rather than experience risks handing AI companies a convenient moral cover story for continuing to scale the technology regardless of the human cost.
Whichever side turns out right, only one of them stands to profit from the answer.