Credible Theories Sign in

The record

Giulio Tononi proposed integrated information theory in 2004. It says consciousness is tied to how much a system's parts act together as one unified whole, a quantity called phi. On this view, what matters is the structure of cause and effect, not what the system is made of.

Rival theories exist, such as global workspace theory. In 2023 a large collaboration reported the first results of an adversarial test designed to pit them against each other. Neither theory came out clearly ahead, and the results drew open criticism, including a public letter by many researchers calling IIT pseudoscience.

That dispute is why this belongs in the Almost credible group: it is real, funded, published research with real data, but nothing is settled and serious scientists disagree about whether it is even testable.

Questions people ask about Integrated Information Theory

Does integrated information theory say AI can be conscious?

Its authors say no for ordinary computers. A 2024 preprint by Tononi, Christof Koch and colleagues argued that a machine built from CPUs, clocks and separate memory can copy a system's behavior without copying its cause and effect structure, so it would have almost no experience. They left the question open for brain-like or quantum hardware.

Why did scientists call integrated information theory pseudoscience?

In September 2023, 124 researchers posted a letter on PsyArXiv saying IIT should carry that label until its core claims can be tested. Critics also pointed to strange results, such as Scott Aaronson's argument that a large grid of idle logic gates would score as more conscious than a person. Tononi accepted that conclusion rather than dropping the theory.

Who ran the 2023 test of consciousness theories, and what did it find?

The Cogitate Consortium ran it, with more than six million dollars from the Templeton World Charity Foundation announced in 2019. Labs agreed on predictions before collecting brain data. In the first results, two of three IIT predictions passed the preset bar and none from global workspace theory did. The full peer-reviewed paper appeared in Nature on April 30, 2025.

The open question

Still strange: More than a hundred researchers signed a letter in 2023 arguing the theory should be labeled pseudoscience.

The Simulation Argument →
  1. Step 1 of 3 · This fileIntegrated Information Theory
  2. Step 2 of 3 · Next fileThe Simulation Argument
  3. Step 3 of 3 · Next fileMany-Worlds Interpretation
About these videos

Found by a YouTube search for this topic, not checked by us. The player loads from YouTube only when you press play.

Where this comes from

EvidenceAlmost credible

Comes from Tononi, "An information integration theory of consciousness," BMC Neuroscience 5:42 (2004). A research article: An article written by a researcher and published in a scholarly journal or book. Most such journals have other experts check a paper before it is printed, so it carries more weight than a news story, though later studies can still change the picture.

What this means

Real research by named scientists or philosophers, with no proof. This page shows what supports the idea, what is missing and who disagrees.

For Integrated Information Theory, there is real evidence, but it does not settle the question (4 sources listed at the bottom of this page). The main doubt: In 2023 a letter signed by more than a hundred researchers argued that IIT should be labeled pseudoscience.

Serious people have done real work and there is real evidence for the idea, but it falls short of proof. We show what supports it, what is missing and who disagrees.

See all five labels

Sources

  1. A research articleAn article written by a researcher and published in a scholarly journal or book. Most such journals have other experts check a paper before it is printed, so it carries more weight than a news story, though later studies can still change the picture.Tononi, "An information integration theory of consciousness," BMC Neuroscience 5:42 (2004)Primary text
  2. Stanford Encyclopedia of PhilosophyA free online encyclopedia of philosophy from Stanford University. Each article is written by an expert, checked by other experts and kept up to date, which makes it one of the most trusted sources in its field.Stanford Encyclopedia of Philosophy, "Consciousness"Reference
  3. WikipediaA free online encyclopedia written and edited by volunteers. It is a good starting point and usually lists its own sources, but anyone can change it, so we treat it as a guide to better sources, not the final word.Wikipedia, "Integrated information theory"Reference Archived copy, 2026-10-01
  4. Stephen Fleming et al. (124 signatories), 'The Integrated Information Theory of Consciousness as Pseudoscience', PsyArXiv, September 16, 2023Primary text

Quotations come from public-domain translations. A book in our Library links to its Library page; other links open the free online copy or the reference page.

Updated

Credible or not?

Tap one to see how other readers voted.

Discussion