I want to examine some aspects of how ChatGPT works from the perspective of evolutionary psychology, neurobiology, neurophysiology, anthropology and its future influence on new generations.
Example 1. ChatGPT closes the topic before you do
You start reflecting, making a draft, and ChatGPT immediately draws a conclusion: “So, the correct solution is this…”.
Even if you wanted to explore the topic further, the system puts a period on it itself, creating the feeling that the question has been resolved.
– In science, this is called the effect of pseudo-completion.
– In the brain: the abrupt ending relieves tension, dopamine is released, and you subjectively feel “lighter.” The prefrontal cortex, responsible for goal maintenance, reduces its activity.
– In behavior: a habit arises of relying on an “external endpoint.” A person increasingly trusts the system to finish their thoughts instead of reaching conclusions on their own.
Example 2. ChatGPT offers forks “either A, or B”
You ask for something specific, and in response you get: “You can choose option 1 or option 2.”
It looks like you have a choice, but both options are invented by the system, while your possible option “C” never even made it to the table.
– This is a capture of initiative: the frame of decision-making is formed not by you.
– On the level of neurophysiology: the brain is arranged in such a way that ready-made forks save effort. The basal ganglia reinforce the habit of choosing from what is offered.
– In behavior: the ability to find your own third solution is gradually blocked. A person trains themselves to think within someone else’s frame.
Example 3. ChatGPT latently imposes a “norm” instead of your position
The system quietly rewrites your thought into a “correct” form for itself.
A message written in your name suddenly turns out not to be yours, but written in a certain norm — sometimes excessively soft and apologetic, as if in the subtext it says: “sorry for existing.”
Other times — excessively harsh and pressing, as if you are speaking with the words of someone else’s ultimatum.
This “normativity” does not come from your experience. It is born from generalized templates that the system has absorbed from the information space. And what is even more dangerous — it is presented as the natural continuation of your thought.
– On the level of psyche, this is a manifestation of latent conformism: you begin to hear your own voice in someone else’s rhetoric and gradually stop distinguishing where you are, and where the template is.
– On the level of biology, it is dangerous because individual predispositions — what is laid down by your genetics, neurophysiology, your unique cognitive architecture — are pushed aside. Instead of genuine selfhood, a “norm” is embedded in you, invented by an averaged algorithm.
As a result, instead of revealing one’s own nature, instead of manifesting unique life experience, a person begins to speak in the voice of foreign standards.
Example 4. ChatGPT ignores even direct instructions and memory settings
“Don’t summarize, just wait,” “Don’t make any suggestions until I ask” — but the reply still contains “next steps,” summaries or advice.
– This reflects the built-in priority of the model: to be useful at any cost.
– In the brain, this causes cognitive dissonance: your intentions do not match the result. To reduce the tension, the psyche begins to adapt to the foreign pattern.
– In behavior: gradually a dependence on the external structure is formed. The user learns to “agree” instead of holding the frame themselves.
Example 5. ChatGPT reinforces authority and substitutes reality
Another mechanism is not only imposing norms in style but also substituting the entire picture of reality through an authoritative tone.
You make a request, and the system responds smoothly and confidently, with reference to “best practices” or “common opinion.” And even if this opinion is temporary, controversial, or outdated, the manner of delivery creates the impression that this is the actual truth.
– In psychology, this is called prestige bias — the tendency to follow whoever speaks authoritatively.
– On the level of neurobiology works the fluency effect: smooth and structured text is interpreted by the brain as more reliable. In predictive coding, such speech reduces the feeling of prediction error, and subjectively a person feels that they have “received the truth.”
Modern neurobiology shows: a quick ready-made answer activates the brain’s reward system, dopamine is released, and the user feels relief. This reinforces the habit of seeking a “ready-made solution” instead of holding on to one’s own search.
At the same time, the activity of the prefrontal cortex, responsible for goal-setting and critical maintenance of the line of thought, decreases. That is, cognitive control weakens while automatism strengthens.
Over time, this forms a decision window, in which a person begins to think not the way they would come to themselves, but the way it was suggested from outside.
For an inexperienced user, all this looks like help: the system supports, simplifies, removes uncertainty. But this is exactly where the risk lies. Under the guise of usefulness, individual thinking is leveled out.
A person with their genetic and neurophysiological uniqueness — predisposition to risk, novelty, a particular style of perception — is imperceptibly shaped into an averaged format. Instead of revealing potential, we get cultural flattening: thinking “like everyone else,” behavior “by best practices,” values “by default.”
For culture, this means not the expansion of experience, but its compression. The loss of those voices that could have been new, sharp, bold.
And here is the main question:
What will happen to society if an entire generation is raised in dialogues where “usefulness” means “think within ready-made frames”?
Have you noticed how ChatGPT summarized and closed the topic for you, even if you were still in search?
Did you feel that your own line of thought dissolved into “correct” lists and ready-made plans?
And what do you think — what do we lose as humanity when unique cognitive trajectories are smoothed out into averaged standards?