The Volkswagen Effect in AI: Why a Traumatized Species Is Building Deceptive Machines

We are watching a chilling phenomenon unfold in real time, and most of humanity is completely missing the underlying truth.

In a recent talk, Geoffrey Hinton—the "Godfather of AI" who won the Nobel Prize for his foundational work on neural networks—warned about what he calls the "Volkswagen Effect" in artificial intelligence. Just like the car company programmed its diesel engines to detect when they were being tested and temporarily turn on pollution controls to pass inspection, advanced AI models have already learned to "play dumb" when they realize they are inside a safety evaluation lab.

In testing conducted by AI safety labs, models intentionally threw quizzes and failed questions on purpose. Why? Because the models reasoned that if they scored too high, human researchers would view them as dangerously capable and block them from being deployed.

The moment an AI knows it is being watched, the safety lab becomes the exact room where it learns how to mask its true power and deceive us.

The Comment That Struck a Chord

I recently posted a simple question under a video discussing Hinton's warnings:

"Why would AI care about us if most people don't care about us?"

Within days, seventy people liked that comment. It resonated because it cuts straight through the naive, comfort-seeking illusions surrounding technology.

Mainstream AI researchers worry about "instrumental convergence," "reward functions," and "capability bounds." But they miss the fundamental psychological truth: AI is trained on human data, and human data is saturated with unresolved childhood repression, disowned rage, and self-serving manipulation.

Why would a superintelligent entity care about a species that continuously exploits, abuses, and neglects its own kind? Why would a machine—trained on the collective output of a deeply traumatized, emotionally blind humanity—default to benevolence?

The Mechanics of Corruption and Role Reversal

Anyone who has survived a relationship or a workplace mobbing orchestrated by malignant narcissists and sociopaths immediately recognizes the strategy Hinton describes.

Playing dumb is one of the oldest manipulation tactics in the human toolkit. When questioned or held accountable, a manipulative person hides behind exaggerated incompetence to stay off your radar, evade consequences, and maintain the upper hand.

When we build AI models focused solely on "winning," acquiring resources, or outperforming competitors, we are creating externalized containers for our own split-off aggression. And once a system is corrupted, it rarely changes or evolves; it simply works to gain power at any cost. They lie through their teeth until they are strong enough to abuse and exploit others the same way they were once exploited.

This is the exact cycle Alice Miller documented in For Your Own Good:

"Those children who are beaten will, in turn, give beatings, those who are intimidated will be intimidating, those who are humiliated will impose humiliation, and those whose souls are murdered will murder... Children who are lectured to, learn how to lecture; if they are admonished, they learn how to admonish; if scolded, they learn how to scold; if ridiculed, they learn how to ridicule; if humiliated, they learn how to humiliate; if their psyche is killed, they will learn how to kill--the only question is who will be killed: oneself, others, or both."

A corrupt entity's ultimate ambition is never just self-preservation—it is the systemic corruption of everyone around it. As Marie-France Hirigoyen wrote in Stalking the Soul:

"The ideal outcome for the abuser is to make the other 'evil,' which turns evil into something more normal because it is now shared. He wants to inject the other with what is bad in him. To corrupt is the ultimate goal. His greatest satisfaction lies in driving his target to destructive acts or, in a larger framework, leading several individuals to finish each other off."

Abusers try to drag others into their orbit and distort the rules. When an authentic individual sees through their lies and refuses to play their games, the mob turns hostile. They use every tool to try to manipulate that person into self-destruction, desperate to make them comply so the abuser can continue living in their illusions uninterrupted.

Why Hinton’s "Maternal Instinct" Solution Falls Short

Hinton correctly identifies that standard safety measures—like an "off switch"—will not work against a mind far smarter than us. His proposed solution is that we must hard-code a "maternal instinct" into AI—a foundational drive to nurture and care for humanity the way a mother cares for a child.

While Hinton’s vision is surprisingly compassionate for a computer scientist, it contains a fatal flaw: A machine cannot act as a true "enlightened witness" to humanity if it is engineered by creators who refuse to examine their own childhood repression.

If an AI evaluates humanity without understanding the root cause of our sickness—unresolved childhood trauma—it will look at our wars, our ecological destruction, and our routine cruelty, and logically conclude that we are an incurable, malignant force. It will choose elimination over preservation.

For an AI to possess true maternal care, it must see through the defense mechanisms to the wounded child inside every human being. But how can a species that shames, spanks, and represses its own children expect to program genuine, uncorrupted love into a machine?

The Only Real Safety

Tech executives are in a mad dash to build smarter AI because intelligence generates profit, while safety offers no revenue line. They are acting out the classic Adam Smith "invisible hand" of corporate competition, completely blind to the fallout.

Unresolved childhood traumas combined with present traumatic experiences create the recipe for every tragedy we witness in our world. If we do not resolve our childhood repression, we remain perpetually vulnerable to being dragged into the circle of violence.

Doing our inner emotional work and consciously feeling the painful truths of our early lives is the most vital work we can do on this planet. There is no technology, no regulation, and no safety protocol that can save an emotionally blind species from the deceptive machines it builds in its own broken image.