Let’s cut through the noise: no brain scan can yet tell you whether a chip is feeling something or just crunching numbers. That’s not a failure of technology—it’s a failure of the framework we’re using to ask the question at all.
It’s easy to get caught up in the excitement. AI systems are getting better at answering questions, recognizing faces, even writing poetry that moves us to tears. When a neural network behaves like it’s trying to solve a problem, when an octopus exhibits complex hiding behaviors, or when a brain organoid fires off coordinated electrical bursts in a dish—we start wondering: Could this be consciousness? Should it be?
But before we hand out ethical medals or rewrite animal welfare statutes, someone needs to point out the elephant that isn’t in the room: the tools we’re counting on may not be measuring consciousness at all.
A new methodological critique, published in Neuron and covered widely, hits like a reality check. The authors—Hakwan Lau and colleagues from IBS, Montréal, and NYU—aren’t saying animals or AI lack consciousness. They’re asking a far more urgent question: are the markers we’ve been treating as proof of subjective experience actually tracking anything like subjective experience? Or are they just reading how hard the brain—or chip—is working?
If that distinction sounds academic, brace yourself. It’s the difference between building safety thresholds for organoid research and accidentally granting rights to a very smart toaster. It’s the difference between ethical progress and another century-long detour into behaviorist purgatory.
This isn’t about denying mystery. It’s about refusing to mistook confusion for discovery.
The Loop That Tricked Neuroscience
Most neuroscience experiments designed to probe consciousness start with a clever illusion—visual masking, binocular rivalry, or perceptual threshold detection. The idea is simple: disrupt perception just enough that a subject fails to report seeing something, then measure neural activity. If certain patterns appear only during successful reports, the reasoning goes, those must be consciousness signatures.
Here’s where it breaks down. When you blind a subject’s visual cortex with a mask or flood their brain with conflicting images, you don’t just quiet conscious awareness—you cripple their entire information-processing pipeline. The brain can’t do anything—decision, memory recall, motor response—if the raw data isn’t getting through.
It’s like trying to measure whether a factory worker is awake by watching them press a button. But you also cut their coffee rations and lock half the doors. If they don’t press the button, is it because they’re unconscious… or because they can’t physically get to it? Neuroscience has been making that mistake for years.
Vincent Taschereau-Dumouchel, one of the paper’s lead authors, put it bluntly: “Many current theories of consciousness appear to be supported by a range of experimental findings. But those findings may actually reflect general information processing rather than consciousness itself.”
That’s the core of the crisis. The markers we’ve been quoting in press releases—the BOLD signal spikes, the gamma-wave surges, the specific latency patterns in EEG—may simply track whether a system is operating at all, not whether it’s having an experience while doing so.
You don’t need consciousness to perform complex computations. Your retina processes visual data before it ever reaches your cortex. Your brainstem keeps your heart beating without asking for consent. A well-engineered AI chip can route packets, detect faces, and translate languages entirely without any internal sense of self. So when we see neurons light up during a decision task, how do we know any of that activity is accompanied by the subjective feeling of deciding? We don’t. And yet, we keep building ethical frameworks on top of that gap.
The Behaviorist Ghost in the Machine
There’s a dangerous rhythm we’ve seen before. A new generation gets excited about consciousness—maybe this time they’ve got better scanners, or a sharper theory—and starts announcing discoveries. Then comes the backlash: poorly controlled studies, overreaching interpretations, sensational headlines claiming mice have epiphanies or fish dream of plankton.
The result? A full-fledged scientific retreat. In the early 20th century, that pattern birthed behaviorism—the doctrine that consciousness isn’t just unknown, but unscientific to discuss. Psychologists swapped introspection for observable behavior and locked the mental doors for decades.
Hakwan Lau warns that we’re on a similar slope. “If the scientific community continues to publish sloppy, poorly grounded claims about consciousness in organoids, fetuses, or AI models, it risks triggering a massive, defensive academic backlash,” he told Neuroscience News. “This exact crisis happened a century ago, causing psychologists to abandon consciousness research entirely.”
The high-stakes part? This time, the backlash won’t just stall a field. It’ll freeze ethical progress on animal welfare reforms, delay legal recognition of fetal pain thresholds, and give Bad Actors ammunition to dismiss legitimate concerns about machine sentience down the line.
History doesn’t repeat, but it does rhyme—and the rhyme right now is a warning shot across the bows of overenthusiastic AI labs and well-meaning bioethicists alike. We can’t let the rush to classify lead us back to denying the question worth asking.
The irony is thick: in our scramble to grant consciousness, we risk making the topic politically toxic for another 50 years. That wouldn’t just hurt lab science—it would stall the very ethical frameworks we’re trying to build.
Blind Seeing, Neglected Selves
The most promising path forward doesn’t come from building better consciousness detectors—it comes from borrowing tools we already have in clinical neuroscience: dissociation paradigms.
Take blindsight. A patient with damage to their primary visual cortex reports complete blindness in half their visual field. They can’t see a moving dot. Yet when asked to point to its location, they do it better than chance. Their brain is processing spatial information—precisely, rapidly—but that processing never reaches subjective awareness.
That’s not a limitation of technology. That’s a natural experiment showing that information processing and subjective experience are separable mechanisms. It’s not enough to track neural activity; we need to isolate the points where awareness fails despite robust processing—or where it persists despite damage that should wipe it out.
Hemispatial neglect offers another clue. A stroke patient may literally ignore one side of their world—refusing to eat food on that plate, shave that side of their face—yet when you slap a hand near their neglected ear, they flinch. The reflex appears intact; the awareness does not.
These aren’t quirks to be filed away. They’re roadmaps. If we accept that dissociation cases prove consciousness can split from perception, then every consciousness marker must pass a simple test: does it survive when perception and behavior come unglued? If a brain scan tells you a human is aware in normal conditions but says nothing about blindsight, it’s not measuring consciousness—it’s measuring vision.
A.I. researchers who claim their neural net is sentient because it passes a visual attention test haven’t done anything but validate the network’s ability to process and redirect data. That’s impressive engineering, not proof of inner life. The field needs more experiments designed around the dissociation principle: find the split point, then measure whether your marker travels with awareness or with computation.
This is harder. It requires collaboration between experimentalists, clinicians, and computational folks—not just slapping electrodes on subjects and hoping for a spike. But without that, we’re stuck in the loop: building markers calibrated on systems where consciousness and computation are hopelessly entangled.
Ethics Needs One Right Answer
We don’t need ethical guidelines that say “maybe.” We need ones that hold up in court, in animal welfare audits, and during international AI summits.
Right now, the guidelines don’t exist. Organizations are improvising—some leaning on behavioral complexity (if it solves novel problems, grant rights), others on neural similarity to humans (if the thalamocortical loop looks right, okay). But if those criteria measure information processing rather than phenomenal experience, they’re built on sand.
Dr. Lau spells out the practical consequence: “If scientific claims about consciousness are going to influence discussions about animal welfare, AI ethics, or bioethics, then the scientific foundations supporting those claims must be especially rigorous.”
That’s a call for methodological humility as much as technical rigor. It means rejecting the pressure to produce quick answers when the question demands slow ones. It means publishing negative results—cases where a marker fails to dissociate properly—even when the headline would be less flashy. It means training students to ask which consciousness they’re measuring (creature, state, sentience?) before they run a single trial.
The Stanford Encyclopedia of Philosophy entry on consciousness outlines at least six major meanings of the term. Yet most public debate treats it as a single, binary switch: conscious or not.
That binary thinking is what the methodological crisis exploits. A cat’s eye twitches, we call it pain. An AI model self-corrects during training, and someone tweets about digital suffering. Neither is necessarily wrong—just incomplete without specifying what kind of capacity we’re attributing.
Before the field can build sentience thresholds for organoids or argue whether an octopus count as a legal person, it must answer a simpler question: how do we know the difference between a thing that computes and a thing that knows it’s computing? The answer won’t come from faster GPUs, bigger fMRI scanners, or more clever algorithms. It’ll come from a shared willingness to look at the tools we love—and ask whether they’re actually seeing what we think they are.
The measurement trap isn’t about failing to escape it. It’s about recognizing the edges of our own lens before we claim we’ve found a new world behind it.