Philosopher Shannon Vallor and I are in the British Library in London, home to 170 million items—books, recordings, newspapers, manuscripts, maps. In other words, we’re talking in the kind of place where today’s artificial intelligence chatbots like ChatGPT come to feed.

Sitting on the library’s café balcony, we are literally in the shadow of the Crick Institute, the biomedical research hub where the innermost mechanisms of the human body are studied. If we were to throw a stone from here across St. Pancras railway station, we might hit the London headquarters of Google, the company for which Vallor worked as an AI ethicist before moving to Scotland to head the Center for Technomoral Futures at the University of Edinburgh.

Here, wedged between the mysteries of the human, the embedded cognitive riches of human language, and the brash swagger of commercial AI, Vallor is helping me make sense of it all. Will AI solve all our problems, or will it make us obsolete, perhaps to the point of extinction? Both possibilities have engendered hyperventilating headlines. Vallor has little time for either.

She acknowledges the tremendous potential of AI to be both beneficial and destructive, but she thinks the real danger lies elsewhere. As she explains in her 2024 book The AI Mirror, both the starry-eyed notion that AI thinks like us and the paranoid fantasy that it will manifest as a malevolent dictator, assert a fictitious kinship with humans at the cost of creating a naïve and toxic view of how our own minds work. It’s a view that could encourage us to relinquish our agency and forego our wisdom in deference to the machines.

It’s easy to assert kinship between machines and humans when humans are seen as mindless machines.

Reading I was struck by Vallor’s determination to probe more deeply than the usual litany of concerns about AI: privacy, misinformation, and