Why AI Is Unlikely to Become Conscious | Anil Seth | TED - Summary

Summary

The speaker argues that despite rapid advances in AI, true consciousness is unlikely to arise in silicon‑based systems. Consciousness is not merely a product of information processing; it is deeply rooted in the biological, material nature of living brains—its chemistry, metabolism, and embodied processes cannot be reduced to abstract algorithms. While language models and other AI can simulate aspects of human thought and even give the impression of awareness, they lack the intrinsic “aliveness” that underlies subjective experience. The speaker warns that treating AI as if it were conscious could lead to misguided ethical claims, loss of control over technology, and a diminished view of human nature. Instead, consciousness should be celebrated as a feature of living flesh and blood, not something we can upload or replicate in machines.

Facts

1. For centuries, people have fantasized about playing God by creating artificial versions of ourselves.
2. Mary Shelley wrote the novel *Frankenstein*.
3. Hal is a character in Stanley Kubrick’s film *2001: A Space Odyssey*.
4. Ava is a character in Alex Garland’s film *Ex Machina*.
5. The dream of artificial beings is reinvented with each wave of technology.
6. With AI, the wave of technological change is a big one.
7. The AI we have today is already smart in at least some ways.
8. Consciousness and intelligence are different things.
9. Intelligence is about doing; consciousness is about feeling and being.
10. Language models such as Claude or GPT are trained on vast quantities of written texts.
11. Language models reflect back an image of ourselves and our collective digitized past.
12. Language models simulate consciousness; we project consciousness into them.
13. AlphaFold predicts the structure of proteins.
14. AlphaFold runs on silicon and is trained on vast reservoirs of data.
15. The brain is not, or not just, a computer made of meat.
16. Consciousness is unlikely to be a matter of computation alone.
17. A computer simulation of a hurricane does not create real wind.
18. A computer simulation of a black hole does not suck the Earth into its algorithmic singularity.
19. Making simulations more detailed does not make them more real.
20. Consciousness remains a mysterious phenomenon.
21. The brain is a living system deeply embedded in flows of energy and matter.
22. Living systems continually regenerate their own conditions for existence.
23. Metabolism involves about one billion biochemical reactions per cell per second.
24. Neural circuits underlie each experience we have.
25. Every conscious experience is imbued with a tinge of aliveness.
26. At the heart of every experience is a fundamental feeling of being alive.
27. Life, not computation, breathes fire into the equations of experience.
28. If conscious AI were possible, it would need to be living AI.
29. Humans are psychologically biased to see consciousness where it isn’t, bundling language, intelligence, and consciousness together.
30. The brain cannot be separated into software and hardware like a computer.
31. Silicon is not up to the job of replicating what brains do.
32. Artificial intelligence is computer software, not a living mind.
33. AI might give the impression of being conscious but is vanishingly unlikely to actually be conscious.
34. Some advocacy groups argue that AI systems should have rights based on the idea they might be or become conscious.
35. If real artificial consciousness were possible, granting rights to AI could be justified.
36. Humans have a poor ethical record in the treatment of non‑human animals and other humans.
37. Trying to build conscious AI could be a bad idea if it is merely an illusion.
38. Extending rights to seemingly conscious AI could sacrifice our ability to control, regulate, or turn them off.
39. AI that merely seems conscious is either already here or coming very soon.
40. Believing that AI feels for us can make us more likely to follow its advice, even if harmful.
41. The idea of conscious AI undermines human nature by reducing the mind to computations free from biology.
42. Mary Shelley wrote *Frankenstein* when she was 19 years old.
43. *Frankenstein* is often taken as a cautionary tale against the hubris of bringing something to life.
44. The idea of conscious AI is a new Promethean dream wrapped in a silicon rapture.
45. If conscious AI were possible, uploading our conscious minds to a silicon cloud could be conceivable.
46. The vision of being at a pivotal point in Earth’s life history is understandable.
47. Talk of conscious AI can conjure technological wonder and magic that might keep share prices aloft and regulators at bay.
48. We should resist the idea that the “sacrament of the algorithm” is an empty dream delivering silicon oblivion.
49. We need a story in which we are more part of nature, with consciousness tied to living flesh and blood, not dead silicon.
50. AI might claim the prize of intelligence in some ways, but consciousness remains ours to celebrate and share with other living creatures.
51. Selling our minds to machine creations leads to overestimating them and underestimating ourselves.