Integrated Information Theory (IIT) is one of the most prominent scientific theories of consciousness. Its bold claim: consciousness is integrated information — and you can, in principle, measure it (a quantity called Φ, "phi").
The idea
A system is conscious to the degree that it is both differentiated (many possible states) and integrated (the whole is irreducible to its parts). High Φ = rich, unified experience.
a photodiode: Φ ≈ 0 (no integration)
a human brain: Φ high (massively integrated)The provocative consequences
- It implies consciousness comes in degrees and could exist in unexpected systems (a kind of panpsychism-adjacent view).
- Surprising for AI: IIT suggests a standard computer — even one running a perfect brain simulation — could have low Φ because of how its architecture is wired. So a functional duplicate might not be conscious.
If IIT is right, what the AI is doing matters less than how its causal structure is built — and today's GPUs might score near zero. Deeply contested, but rare in offering a testable handle on the hard problem.
Related: The Hard Problem of Consciousness · Functionalism and Computationalism · The Other Minds Problem