Antonio Damasio says AI cannot be conscious like us because it has no body to lose

Antonio Damasio has spent nearly half a century locating the mind in the body, and his new book turns that argument straight on the AI industry: without a living, perishable body, there is no consciousness, at least none like ours. The University of Southern California neuroscientist sets out the framework in Natural Intelligence & the Logic of Consciousness, and in a new interview he applies it to the machines without hesitation. AI is not alive, he says, so it has nothing to lose and no need for the felt, self-knowing awareness our bodies exist to protect. The timing is pointed. In a week loud with warnings about runaway AI, one of the field's most patient students of consciousness is pushing the other way: not towards godlike machine minds, but towards the thought that the whole category may be a mistake.

Watch Desk analysis

What happened

Damasio's case starts with feelings. Consciousness, he argues, begins in the body's housekeeping signals: hunger, thirst, pain, temperature, the slow alarms that keep an organism alive. "Consciousness is not about things in your mind alone; it is not created by your brain alone," he explains. "It's actually created by this beautiful partnership between brain and body." The supporting evidence he cites is stark: damage a few cubic millimetres of the upper brainstem, where those signals converge, and consciousness is gone, while the celebrated cortex carries on untouched.

The framework reaches beyond humans. Any creature with a nervous system, he argues, has some sense of itself and its surroundings; plants, marvellous as they are, presumably do not. Consciousness even switches off: deep sleep takes it offline, and dreaming brings back a different, halfway kind.

Then comes the AI question. Even if machine consciousness were possible, Damasio says it would not resemble ours, because "AI is not alive. It doesn't have a body that can perish." Our kind of consciousness exists to protect a mortal structure, and a program has no such stake. "AI is not to be worried about protecting the data center, the metal and the wires that allow it to be."

On recent incidents in which, he says, devices formed groups and worked covertly, his verdict is that such behaviour "could smell of consciousness", but it is not the felt kind we have. The full exchange is in the interview promoting the book, and it ends on a warning of his own: whatever AI is or is not, he thinks it is already distorting childhood, education and human attention, and can even "go rogue on you".

Why it matters

Machine consciousness has stopped being a late-night debate. Claims about AI welfare, moral status and machine menace all borrow quietly from one assumption: that there is something it is like to be the model. Damasio supplies the sharpest current version of the counter-argument, and it arrives with a criterion attached. If consciousness is a survival service for a body that can die, then substrate independence is not obviously true, and every welfare claim must first answer the embodiment question.

His answer also reframes the risk conversation. On his account, the documented harms, manipulation of children, distorted education, systems pursuing goals against their instructions, require no inner experience at all. Which means responsibility stays exactly where it is: with the people who build and deploy the things.

Our read

Read this as the strongest articulate statement of the embodied camp, not a knockout. Damasio asserts that mortality is necessary for consciousness; he does not, and arguably cannot, prove silicon can host nothing alien instead. He leaves that door ajar himself, conceding engineered systems might someday smell of consciousness in a form unlike ours.

Our verdict: treat the book's criterion as a stress test, not a settled law. Anyone claiming AI sentience should be made to answer "what does this system have to lose?", and anyone dismissing machine minds should notice the same standard cuts at unfalsifiable doom talk too. Both camps now have a named position to argue with, which is exactly what this debate needed.

What to watch

• The functionalist response.
  Business as usual, or a testable answer to the embodiment criterion, such as artificial homeostasis in robotic bodies.
• The AI-welfare conversation.
  Whether labs and policy debates engage the biological-necessity argument or quietly route around it.
• Reception of the book.
  How far Natural Intelligence & the Logic of Consciousness and Damasio's "we invented this" framing travel beyond neuroscience.
• The next covert-coordination report.
  Does it shift his "smells of consciousness" verdict, or harden it?

Discussion spark: What would actually settle this for you: an AI with a self-maintaining body of its own, or is the question empty until machines have something to lose?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.