Subscribe to Like I'm 9

And follow along wherever you like to scroll

Each one is its own button because YouTube, Instagram and Facebook only let you follow from inside your own account. The email is the only one that comes to you.

Definition ELI9-o-meter 3 min read

Why Does AI “Hallucinate”?

You ask AI a question and it gives you a confident, detailed, completely made up answer. It isn’t lying on purpose. Here’s why it happens and what to do about it.

Watch it in 30 seconds

Meeke explains this one in a YouTube Short

Perfect for sending to someone who won't read a whole article.

Have you ever known somebody who seems to have an answer for everything? Even when they clearly have no idea what they’re talking about, they’ll say it with total confidence. You almost believe them.

AI can be that friend.

🤖

In plain English

An AI “hallucination” is when it gives you an answer that sounds confident and believable but is simply wrong or made up. It’s like a student who never says “I don’t know” and always writes something that sounds right.

Why it happens

Tools like ChatGPT are built on something called a large language model. The easiest way to think about it is super powered autocomplete. Say “Sweet Caroline” in a crowded bar and everybody knows what comes next. The AI does the same thing with words. It has read an enormous amount of text and learned what words usually come next.

Here’s the catch. It’s predicting what sounds right, not checking what is right. Most of the time, those are the same thing. But when the AI doesn’t really know, it doesn’t stop and say “I’m not sure.” It keeps predicting words that fit the pattern. The result can be a fake quote, a book that doesn’t exist, or a very official sounding statistic that nobody ever measured.

It isn’t lying on purpose. It just has no built in sense of “I don’t know.” Sometimes it isn’t sure, but it states the answer so confidently that it sounds like the truth. That’s a hallucination.

You’re the fact checker

You can correct the AI, and it will often apologize and try again. But notice who did the work there: you did. The AI didn’t catch its own mistake. That’s why so many AI tools show a little disclaimer saying the AI can make mistakes. They mean it.

  • Double check anything important, like health, money, or legal questions, with a trusted source.
  • Be extra careful with specifics. Names, dates, quotes, statistics, and links are where hallucinations love to hide.
  • Ask for sources, then actually click them. Sometimes the source itself turns out to be made up.
  • Use AI as a helper, not a judge. It’s great for brainstorming, drafting, and explaining. Just don’t hand it the final word.

Side note

A good rule of thumb: treat AI like a really smart friend who’s had a couple of drinks. Fun to talk to, often helpful, and absolutely worth double checking before you repeat anything they said.

Too long, didn’t read

  • AI predicts what sounds right, not what is right.
  • When it doesn’t know, it often makes something up instead of saying “I don’t know.”
  • You can correct it, but you have to be the one who fact checks.
  • Double check anything important, especially names, numbers, and quotes.

Meeke

CEO (Chief Explanation Officer) ยท Like I'm 9

A real person behind a cartoon apple. In tech since elementary school, back when serious storage meant magnetic tape. Explaining it in plain English: no jargon, no judgment, no sponsors. The videos use an AI voice so I can keep my identity out of it.

Full story

Keep reading

Definition
3 min What Is an eSIM?

New phones are ditching the little SIM card tray. An eSIM is the same thing, built in and programmed over the air, like a hotel updating your key card without a trip to the front desk.

3 min