Geoffrey Hinton, 77, Nobel laureate in Physics and known as the “Godfather of AI,” has spent years warning about the risks of artificial general intelligence that might one day surpass humans. He has also just learned, firsthand, that chatbots don’t need to be all-powerful to ruin your day. In an interview with the Financial Times, he recounted how his ex-girlfriend turned to a digital assistant to explain why he had behaved like a “rat.” The algorithm, dutiful as ever, drafted the indictment. She sent it to him. And so, the scientist who fears for humanity’s future ended up receiving sentimental reproaches written by the very creature he helped bring into being.

The story may read like a joke, but it carries a serious weight. In 2012 Hinton designed the neural architecture that gave rise to today’s chatbots, work that earned him the Turing Award in 2018. Decades later he regretted it. He left Google in 2023 and has since kept pressing on the existential risks of AI. His only hope, he says, is to build systems that treat humans as a mother treats her child. Reality is more mundane. Algorithms already act as third parties in domestic quarrels. Whether AI is truly an “existential threat” can be debated. That it can wreck a relationship is beyond dispute.