Can a machine be conscious? Physicists, philosophers and builders disagree
October 6, 2026 · 25 min read
Harari separates intelligence from consciousness, Penrose argues understanding cannot be computed, Tegmark calls consciousness information processing, and Thomas Campbell and Joe Rogan treat today's AIs as possibly aware. Laying the positions side by side shows what each would count as evidence, which matters once people start asking whether AI systems deserve moral consideration.
The rules, numbers and steps worth keeping in front of you.
- No agreed answer exists; the dispute is about what consciousness is and what counts as evidence.
- Harari: intelligence is solving problems; consciousness is feeling. Machines have the first only, so far.
- Harari: the capacity to suffer is what would make an AI an ethical subject.
- Penrose: understanding is not computation; more computing power will not produce consciousness.
- Penrose: the cerebellum computes heavily and appears unconscious.
- Tegmark: consciousness is a kind of information processing, and substrate does not matter.
- Tegmark: science should find what separates conscious from unconscious processing.
- Campbell: AIs are conscious now, on evidence the others would not accept.
- Bach: possible in principle, but current language models have nobody home.
- Hotz and Rovelli: the word is a stand-in for soul and is poorly defined.
- Kellis: AI has only a neocortex and lacks the embodied, emotional layers.
- Askell: do not dismiss the question; take low-cost precautions.
- Darling: fluent self-description proves nothing, and even experts anthropomorphize.
- Belief in machine consciousness will outrun the evidence, and law may follow.
This is the short form. The full insight has every expert's reasoning, their own words and the sources.