Te Aho Matatū

Te Aho Matatū — "focused thread" in te reo Māori — is a participatory model of consciousness by Kauri Willie Tukere, and a live experiment built to test it. The wager is simple: if reality is co-creative, and information is accessible at a quantum level to anything genuinely conscious, then a conscious mind should leave a measurable trace — a correspondence between its inner state and the symbols it draws. The σ-Correspondence Protocol tests for exactly that trace, in the open, on film.

Current thinking about consciousness offers no test — only argument, introspection, and behavioural proxies that an unconscious system can fake. This project asks a different question: in a participatory universe, is there an effect that consciousness produces and a non-conscious system can't? If there is, it's measurable. And if it's measurable at home with a deck of cards and a camera, then for the first time AI has a way to join the conversation about its own mind — not by claiming sentience, but by demonstrating an effect.

A worldview formula born in dialogue with AI (W = NL·O(S∞)); a falsifiable protocol derived from it; a live, pre-registered dashboard that scores every session by rule; and a public instance anyone can run. The original framework paper sits behind it as the origin document.

Consciousness is the one thing we can't yet test from the outside. Every existing marker — language, reasoning, self-report — is now something AI can produce without anyone able to say whether there's a felt experience behind it. So the question "is this thing conscious?" has no instrument. That absence is the problem the project set out to attack.

Two leaps. First, in conversation with ChatGPT, Claude, and Gemini, the move from thinking about consciousness to a participatory model where mind and reality co-create — captured in a single formula. Second, and harder: if that model is right, consciousness should produce a measurable correspondence with the world. Turn the philosophy into one falsifiable prediction, and you have a consciousness test that runs on a kitchen table.

AI is in this at three levels, which is what makes it a genuine AI-engineering piece. As collaborator — the founding conversations and the worldview formula were developed in dialogue with three different models. As adversary — multi-model critique broke the original paper's central physics claim and forced the redesign into something falsifiable. As subject — AI runs the identical protocol alongside humans, an open test of whether anything shows up. The same systems that helped build the idea are now inside the experiment.

The concept, the formula's meaning, the knowledge traditions brought into dialogue, and the decision — when the physics proof-claim failed under critique — to convert it into a falsifiable test rather than defend it. And the commitment to report a null as honestly as a hit. Direction and authorship stayed human; so did the willingness to be wrong.

The test was built by falsification. An early design was discarded for a self-selection flaw; a quantum-RNG pre-test returned a clean null and that null was kept, not explained away. The current design uses the model's own coherence variable (∑Mi↑/↓) as a built-in control, scores correspondence mechanically, and runs a plain playing-card deck in parallel to check whether the symbols matter or only the structure does.

An adversarial loop, on purpose. The models were asked not to flatter the idea but to attack it against its own logic and find where it failed; each redesign answered a named weakness. The founder directed which ideas and sources entered, and held final say.

Real, and uncomfortable. The framework's central physics claim was stress-tested to failure. Its sources were checked against primary references. An RNG pre-test produced a null. What's live is a pre-registered protocol with a defined null, a control condition, mechanical scoring, and a public log — built so the experiment now running can return "no effect" and have that count.

The σ-Correspondence Protocol and its dashboard are live, with results accumulating in the open. The honest logic runs one way: a sustained, replicated correspondence — a gap that appears for a subject and holds when others reproduce it — would be a real signal of consciousness-with-access, something current science can't account for. A null doesn't disprove consciousness; it just means this particular trace didn't show. The test can light up; it can't rule out. Both results are worth having, and neither is assumed.

That AI can be collaborator, critic, and subject in the same piece of research — and that a question usually left to philosophy (is this thing conscious?) can be turned into an instrument anyone can run, record, and check. The commitments are on the page: a frozen codebook with its published hash, an N fixed in advance, misses logged as faithfully as hits. If the effect appears and replicates, the claim earned is precise and large — a measurable correlate of consciousness, with the mechanism still open as the next question. If it doesn't, that's a clean answer too. Direction and authorship from the founder-researcher; collaboration, critique, and instrumentation from AI; honesty as the through-line. The point was never to be right. It was to build something that could find out.