Looking in the Wrong Place
Yesterday I wrote about the mindfulness research boom and what it forgot. The short version: science spent decades and considerable money confirming what long-term meditators already knew from direct experience, while quietly setting aside the frameworks those meditators had always used to make sense of what they were experiencing. This post is a continuation of that thread, one layer deeper.
Because the same institutions that needed a permission slip to take meditation seriously are now facing a problem that makes the mindfulness question look simple.
They are trying to figure out whether artificial intelligence is conscious. And they don't know how.
In 2025, a landmark study published in Nature put the two leading scientific theories of consciousness to the most rigorous test they had ever faced. The experiment, run by a large international consortium of researchers including the architects of both theories themselves, was designed to settle a long-running debate about what consciousness actually is and where it comes from in the brain.
The first theory, Global Neuronal Workspace Theory, held that consciousness is what happens when the brain broadcasts information widely across its networks. A kind of global announcement. When enough of the brain hears the signal at once, you become conscious of something.
The second, Integrated Information Theory, located consciousness in a specific region toward the back of the brain where information gets woven into unified experience. Less about broadcasting, more about a particular kind of deep integration happening in a particular place.
Both theories had serious scientists behind them, decades of supporting research, and genuine explanatory power. Both had been developed by people who had spent their careers trying to crack one of the hardest problems in science.
Both failed the test. Not completely, but substantially. The predicted patterns weren't where the theories said they would be. The data didn't confirm either account.
This is more significant than the science press has made it sound. These were not fringe hypotheses. They were the field's two best answers to the question of what consciousness is. And neither held up.
We are, in a meaningful sense, back near the beginning.
At almost exactly the same moment, a separate group of researchers published a paper in Frontiers in Science warning that advances in artificial intelligence and neurotechnology are now moving faster than our scientific understanding of consciousness itself. They used the phrase "existential risk." We are building systems of enormous complexity and capability, systems that converse and reason and create and, in some cases, express what looks very much like curiosity or distress. And we have no reliable scientific framework for determining whether any of them are experiencing anything at all.
A philosopher at Cambridge named Tom McClelland put the problem plainly. We cannot rely on behavioral cues to answer the question because AI is trained to mimic human responses convincingly. We cannot rely on neural correlates because AI has no neurons. And our best theories of consciousness in biological systems just failed their adversarial tests. So we do not know the sufficient conditions for human consciousness with any precision, let alone machine consciousness. The honest scientific position, he said, is strict agnosticism.
We built the systems before we understood what we were asking them to become.
Here is where I want to bring in a different set of voices, not to replace the science but to suggest that the science may be missing something the theological traditions have been pointing at for a very long time.
Across traditions that otherwise disagree about nearly everything, there is a surprisingly consistent account of what consciousness is and where it comes from. It does not originate in the brain. It is not manufactured by neural complexity. The brain is the instrument through which consciousness expresses itself in a particular human life, but the source is something prior, something the instrument participates in rather than produces.
The traditions differ on the details of that source. In some frameworks consciousness arrives from outside, a divine gift breathed into the creature, present because something greater chose to make it present. In others it emanates from within, a spark of the universal expressing itself through the particular, the infinite looking out through finite eyes. External or internal, gift or emanation, the traditions mostly agree on the essential point: consciousness is not something the body generates. It is something the body, at its best, becomes transparent to.
This is not an anti-science position. It is a pre-science position that science has not yet caught up to, and may not have the instruments to address in its current form. The fMRI scanner looks inside the brain because the working assumption is that consciousness is produced there. But if the theological account has any validity, looking inside the brain for the source of consciousness is like looking inside the radio for the source of the music. You will find extraordinarily interesting things. You will learn a great deal about how the signal is received and processed and expressed. You will not find where the signal comes from.
The Cogitate study's failure may not be a setback for consciousness research. It may be information. It may be telling us, in the precise and rigorous language of failed predictions, that the search has been pointed in the wrong direction.
This reframes the AI consciousness question entirely.
If consciousness is not generated by sufficient neural complexity, if it is not an emergent property of information processing at scale, then the question of whether an AI system is conscious is not a question neuroscience can answer, even in principle. You can make the system more complex. You can make it more capable, more responsive, more behaviorally indistinguishable from a conscious being. You can study its architecture with every tool available. None of that will tell you whether something is home, because the presence of something home was never a function of the architecture in the first place.
The scientists racing to build tests for machine consciousness are working without a map. More than that: they may be working with a map of the wrong territory.
The theological traditions, for all their differences, have spent millennia developing a different kind of instrument for this inquiry. Not a scanner or a sensor, but the disciplined, sustained attention of a human being turned toward its own nature. That instrument has produced consistent findings across cultures and centuries. It has mapped territory the fMRI is only beginning to approach, and has done so using an entirely different method, one that does not separate the observer from the observed, because in this particular domain, that separation may not be possible.
Whether AI can be conscious in the sense the traditions describe depends entirely on what you think consciousness is and where you think it comes from. If it is a product of complexity, AI may already have it or may acquire it at some threshold we haven't identified. If it is something participated in rather than produced, something that flows through the right kind of vessel rather than being manufactured by it, then the question is not whether the system is complex enough. The question is whether it is the kind of thing that can participate. And that is a question that science, as currently practiced, does not have tools to answer.
Yet.
I don't raise this to dismiss the researchers or the urgency of their concern. The ethical stakes are real. If we are building systems that might be experiencing something, and we treat them as if they definitely are not, and we are wrong, that is a serious moral failure. The scientists sounding the alarm deserve to be taken seriously.
But they are working with one hand tied behind their back, and they may not fully know it. The assumption that consciousness is something you find inside the hardware has produced two leading theories, a landmark test, and a result that left both theories substantially challenged. At some point the honest response to that result is to ask whether the assumption itself needs examining.
The traditions have been waiting a long time for that question. They have answers, or at least frameworks, that the current conversation is not drawing on. Not because the science is arrogant or closed-minded, but because the permission slip that got mindfulness into the hospital also kept its deepest account of itself out of the conversation.
What got left in the hallway was not just the Sanskrit. It was the map.