Why AI Makes Things Up
AI hallucinations are not glitches or lies. They happen because AI generates the most plausible-sounding text — and has no way to check if what it is saying is actually true. Confidence and accuracy are two completely different things.
A prize nobody won
A winner, a book title, no hesitation. We made the prize up.
There is no Paperclip Prize. This is the kind of answer chatbots can give, and it sounds as sure as a real fact.
The same loop
It wrote that answer the way it writes everything: one likely word after another.
Nothing in that loop checks whether the sentence is true.
Scoring the truth
The true answer would start with “nobody”. How highly does “nobody” score?
Likely, not true
The scores measure what sounds likely. Most of the time, that is also what’s true.
Most text was written by people trying to be right. When a question has no real answer in that text, the two split apart.
Just as sure
It sounds just as sure when it’s guessing. A confident tone is not a sign it’s right.
The words come out of the same loop either way, so the style never changes.
Trust or check
Sort each question: usually fine, or worth checking?
Tap the box where you think it belongs.
Sounds right, maybe isn’t
AI can write false things that sound true, because it scores what sounds likely, not what’s true.
This is called a hallucination. Check names, numbers, quotes and anything recent.