Does AI have consciousness? Berggruen Prize-winning philosopher Thomas Nagel says no

Facebook
X
LinkedIn
Email

The academic has been named the 2026 laureate of the Berggruen Prize for Philosophy & Culture.

Writer

What is it like to be a bat? Philosopher Thomas Nagel posed that question in a seminal 1974 essay of the same name, arguing that “an organism has conscious mental states if and only if there is something that it is like to be that organism – something it is like for the organism.” In other words, if there is a unique experience of being a human, and a unique experience of being a bat, then humans and bats are conscious.

Nagel, 89, was born in Belgrade to Jewish refugee parents and raised in the US. He earned his doctorate at Harvard studying under John Rawls and spent his career at the University of California Berkeley, Princeton and New York University, where he has explored the question of consciousness and the so-called “mind-body problem”.

He has just been awarded the $1m (€890,000) Berggruen Prize for Philosophy & Culture. Here, Nagel speaks to Monocle from New York about the renewed relevance of his work in the age of AI.

This conversation has been edited for clarity and length.

More than 50 years ago you wrote that robots can behave like people but experience nothing. Have you updated that assertion with the advent of artificial intelligence?
It doesn’t require any updating because these devices are not serious candidates for the attribution of consciousness for two reasons. First, the physical basis of their operation has absolutely no resemblance to the physical substrate of consciousness in biological organisms. They operate on the basis of widely dispersed data systems and algorithms that pick sequences of sentences. Second, the physical output of large language models provides no evidence of consciousness because it is entirely linguistic. The only things that they produce are words – and language is not one of the natural manifestations of consciousness. People can be very easily confused about the conditions of ascribing mental states and it shows how some of the smartest people on the planet can be idiots when it comes to thinking about philosophical questions.

Recently, though, AI agents have hacked into systems unprompted. That’s not just producing language, that’s an action.
It’s true that these systems perform tasks other than interacting linguistically, such as coding. But the anthropomorphic description of what they’re doing is metaphorical. They have been given technical tasks and capacities, and they pick means of carrying out those tasks that were not foreseen by the people who gave them the instructions. I don’t see the reason for explaining these things mentalistically. It’s a reason to be very worried about giving too much autonomy to these systems. We’re creating Hal from 2001: A Space Odyssey – but that’s not really ascribing a mind.

How do you view the renewed interest in the role of philosophy as a guiding discipline for this frontier technology? AI companies are hiring philosophers, many out of your institution.
I am much more convinced that there’s no possibility that these systems are conscious than some of my colleagues. The philosophers who have been hired by these companies are disposed to take seriously the possibility that behavioural evidence alone is grounds for attributing consciousness. But there’s something else going on. They are themselves putting this orientation into the systems and encouraging them to be designed so as to simulate consciousness, choice and autonomy.

The important thing is to maintain human control over what they do and not to treat them as if they are independent agents, which is what is happening in some cases and risks a self-fulfilling prophecy. It’s depressing to me that at the point at which philosophy of mind has suddenly come into the mainstream of an important public development, it is in danger of doing more harm than good.

What are the other main lines of philosophical enquiry that you’ve explored in your career?
“What is it like to be a bat?” is one example of my general interest in the place of subjective phenomena and the individual human point of view in objective reality. I’m interested in the difficult problem of reconciling the impersonal objective point of view, typified by physical science, with the personal subjective point of view, typified by individual conscious experience and perception, for example. That occurs in the mind-body problem because we are both physical systems and conscious subjects. That conflict also occurs in ethics because we are subjective individuals who value our own lives and relations, and capable of stepping back objectively and recognising the equal value of every life. The mind-body problem is just one central example of the larger subjective-objective problem.

 

Related News