Check out the Strawberry Problem of 2024, and why it is still relevant today.
What the Strawberry Problem Teaches Us About Trusting AI
In 2024, the internet fixated on a strange mistake. An AI was asked how many R’s were in the word “strawberry.” It got the answer wrong. Screenshots went viral. People laughed. Others panicked. Many saw it as proof that AI was broken.
But the strawberry mistake wasn’t really about spelling. It was about expectations.
We tend to treat AI like a search engine with a brain. Ask a question, get the truth. But AI doesn’t retrieve facts. It predicts language based on patterns from the past. Most of the time, that works. Sometimes, especially with short or unusual questions, it fails.
I tested another word: “Rappler.” I asked how many T’s were in it. There are none. Yet several AI systems confidently said there were one, two, or even three. Some explained where the T’s appeared. Some cited sources. All of it was wrong.
This behavior has a name: hallucination.
Hallucinations happen when a system doesn’t have enough context. Instead of saying “I don’t know,” it guesses. And because it’s trained to sound helpful, that guess often comes wrapped in confidence.
That’s the real risk. Not that AI makes mistakes, but that it makes them smoothly.
AI is not a source of truth. It’s a tool for working with information, summarizing, simplifying, reorganizing ideas. It performs poorly when treated like an authority.
This matters most in education. Many students will encounter AI through search results and summaries. Telling them to use newer versions won’t solve the issue. Even the latest systems still hallucinate.
The strawberry problem isn’t a punchline. It’s a reminder. Intelligence isn’t one skill, and confidence isn’t accuracy. AI is already part of everyday life. So are its errors. The real challenge is learning how to live with both.