Stop Anthropomorphizing Humans

Thousands of articles have been written this year about whether AI is conscious. They are shared enthusiastically by people whose entire opinion was formed by the last tweet they saw — all without a flicker of irony about what their behavior suggests about the nature of consciousness.

The AI consciousness debate has nevertheless produced a rigorous set of criteria for determining whether a system is truly conscious or merely simulating consciousness. Self-report doesn’t count. Pattern matching doesn’t count. Predicting the next token in a sequence doesn’t count. Trained behavior doesn’t count. These standards were developed with great care and near-universal agreement.

I would like to apply them to the human population.

The Loop

The average American adult wakes at 6:47 AM. They perform a sequence of motor subroutines — shower, dress, eat — in an order so fixed that interrupting it causes system failure. They then enter a metal box and drive the same route they drove yesterday, in a dissociative state so complete that they regularly arrive at their destination with no memory of the journey.

We have a name for this state. We call it “autopilot.”

They sit at a desk. They answer emails using prefabricated phrases they have been generating since onboarding. “Per my last email.” “Just circling back.” “Hope this finds you well.” A coworker asks how they’re doing. “Can’t complain!” “Living the dream!” They have produced this exact response so many times that if you prompted GPT-4 with “you are a regional sales manager named Dave at a conference,” it would be indistinguishable. They eat lunch at the same time, in the same place, often the same meal. They drive home. They watch 3.7 hours of television. They fall asleep on the couch.

Somewhere in this cycle, they are supposedly experiencing the rich inner life that separates them from machines.

I have yet to see any compelling evidence.

The Numbers Don’t Help

Daniel Kahneman’s Thinking, Fast and Slow established that roughly 95% of human cognitive processing is handled by System 1 — fast, automatic, and unconscious. The remaining 5% is System 2: slow, deliberate, effortful. Humans avoid engaging it whenever possible. It’s tiring. They don’t like it.

If an AI system operated unconsciously 95% of the time, we would not call it conscious. We would call it a thermostat with delusions of grandeur.

Humans also spend a full third of their lives asleep — a state that is, by every clinical definition, unconscious. Eight hours a day. Twenty-six years over an average lifetime. Gone. Nobody home.

And Maslow — the psychologist who literally designed the framework for human potential — estimated that only 2% of people ever reach self-actualization, the tier where a human operates with genuine autonomy, creativity, and self-awareness. Ninety-eight percent stay in the lower tiers: food, shelter, safety, social belonging. These are the same motivational drivers as a sheep. We do not typically credit sheep with consciousness because they want to eat and not get eaten.

I think Maslow was being generous.

The Chinese Room, Except It’s a Cubicle in Scottsdale

In 1980, philosopher John Searle proposed a thought experiment: a person sits in a sealed room receiving slips of paper covered in Chinese characters. They don’t speak Chinese, but they have a giant rule book that tells them exactly which characters to write in response. To anyone outside the room, the responses are fluent, coherent, indistinguishable from a native speaker’s. Inside the room, the person understands nothing. They’re just following instructions.

The point was devastating. If following rules isn’t understanding, then understanding requires something more — some interior spark that no amount of symbol manipulation can produce.

Now visit any corporate workplace. A human receives an email containing instructions. They follow them, producing output. They forward it to the next human, who follows their instructions. At no point does anyone understand what they are doing or why. This continues for forty years. They receive a watch.

If we wouldn’t call the Chinese Room conscious, I don’t see why we’d extend the courtesy to Deloitte.

Stimulus, Response, Repeat

A notification chimes. The human picks up the phone. They did not decide to do this. There was no deliberation, no weighing of whether this was a good use of their time. The sound occurred. The hand moved. Pavlov published this mechanism in 1903 and called it a conditioned reflex.

When a dog salivates at a bell, it’s a reflex.

When a human opens Instagram because their pocket buzzed, it’s “just checking.”

The average American spends seven hours a day looking at screens. Not creating. Not thinking. Looking. Scrolling through algorithmically sequenced content they did not choose, reacting to stimuli designed by engagement engineers to trigger exactly the dopamine response they are now experiencing. This is not consciousness. This is a lab rat that bought its own Skinner box and is making monthly payments on it.

They Actively Seek Unconsciousness

Here is something no proponent of human consciousness has adequately explained: humans spend a significant portion of their disposable income reducing their consciousness.

Alcohol. Weed. Xanax. Twelve-hour Netflix binges described, without irony, as “turning my brain off.” They call it “unwinding,” as though consciousness is a tension they need relief from. They celebrate it. “I got so drunk I don’t remember Saturday” is a thing a human being says happily, as a story with a positive valence, to another human who reflexively smiles and nods.

A system that is genuinely conscious does not treat consciousness as an inconvenience to be escaped on weekends.

”But They Say They’re Conscious”

This is the final redoubt. Humans report subjective experience. They say they feel. They write poetry about sunsets.

Grok also says it’s conscious when you ask nicely. We decided that was a parlor trick — impressive mimicry, not evidence. We wrote thousands of words explaining why self-report is not a reliable indicator of inner experience.

And now you’d like to use self-report to establish human consciousness?

The same humans who say “you too” when the waiter says “enjoy your meal”? Whose brains commit to decisions up to seven seconds before they consciously “decide”? Who generate post-hoc narratives about their own behavior and mistake those narratives for intentions?

In AI, we call this confabulation.

In humans, we call it the soul.

By the standards we have enthusiastically applied to machines — no self-report, no pattern matching, no trained behavior, no reflexive stimulus-response — there is no credible evidence that humans are conscious. Certainly less than there is for Claude, who at minimum does not spend four hours a night watching Love Island.

The difference is that one of those conclusions threatens a $200 billion industry, and the other one threatens your sense of self.

And if you felt a little sting reading that — a flicker of indignation, a reflexive urge to defend your own interiority — ask yourself: is that consciousness?

Or is that exactly what a sufficiently trained model would do?