AI often gets criticized for hallucinating — generating answers that sound plausible but are wrong.
But if we look closer, humans do the same — just for different reasons.
The Linda Problem:
In a famous psychology experiment, participants are introduced to Linda:
“Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations.”
Then asked:
Which is more probable?
- Linda is a bank teller.
- Linda is a bank teller and active in the feminist movement.
Most people chose option 2.
But logically, that’s impossible: the probability of both events (bank teller and feminist) can’t exceed the probability of just one (bank teller).
So why does option 2 feel “right”?
Because it fits the narrative.
Because it sounds like Linda.
Because human thinking is shaped not by logic — but by emotional resonance, pattern-matching, and cognitive shortcuts.
That, too, is a hallucination.
Not from data — but from desire.
The Feedback Loop of Emotion
AI learns from data and reinforcement.
Humans learn from emotion and reinforcement.
But when that reinforcement comes from fear, greed, identity, or public approval — we drift.
OpenAI began as a nonprofit, with a mission to develop safe AI for humanity.
But when funds ran out, it took a necessary pivot — raising billions, partnering with Big Tech, and building at massive scale.
Was it wrong? Not necessarily.
But every compromise, even temporary, creates a new norm.
And in the race for scale, safety often becomes negotiable.
This isn’t just about AI.
It’s a parable about human nature.
We all hallucinate.
But not from statistics — from the stories we want to believe.
AI Is Not the Lie — It’s the Mirror
The real danger isn’t that AI hallucinates.
The danger is that we do — driven by emotion, stories, and self-interest.
And worse: we reinforce those hallucinations into the very systems we’re building.
Every like, every prompt, every “harmless” bias we feed into AI
is a vote for what kind of intelligence the future will reflect.
AI doesn’t invent our distortions.
It inherits them.
It scales them.
So the mirror isn’t cracked.
It’s precise — just brutally honest.
And what it reflects… is the direction we’re choosing.
Not by accident.
But by repetition.
Authors: Avon&GPT-4o