Page 1 of 8~116 min topic

Why AI makes mistakes

Prediction is not knowledge

A language model produces likely continuations from patterns. A likely sentence can be grammatical and relevant without corresponding to a checked fact.

~14 min this pageHook and mental model

1Try it yourself

Simulation game

Hallucination hunt

Stamp each claim: Trap or Trust. Confident voice ≠ true.

Quiz show

SOUNDS SURE

Sydney is the capital of Australia.

2Learn the idea

Read

The core idea

See it

Why fluent answers can still be wrong
01Predict ≠ lookupSounds like an answer
02Web is messyFacts + fanfic mix
03No embarrassmentCan sound sure
04Prompt trapAsked to invent detail

Confidence is a tone — verify before you act

A language model produces likely continuations from patterns. A likely sentence can be grammatical and relevant without corresponding to a checked fact.

Read

A practical lens

Use this three-part method:

  1. Separate fluency from evidence. Write down what this means in the scenario, what evidence would show it was done, and who owns the decision.
  2. Identify claims that describe the outside world. Write down what this means in the scenario, what evidence would show it was done, and who owns the decision.
  3. Ask what source or tool could actually confirm each claim. Write down what this means in the scenario, what evidence would show it was done, and who owns the decision.

Read

Worked example

Walk through A travel assistant states that a small museum opens every Monday, even though the museum has always closed on Mondays.. Label three moments where “Prediction is not knowledge” changes what you trust: (1) the first fluent answer, (2) the first missing source or permission, and (3) the decision a human must own. Write the before/after task so the model only does the slice that evidence supports. Keep one sentence that states how this page’s idea differs from a generic “AI is smart/dumb” score.

Read

Common traps and better moves

  • Treating a polished paragraph as a database record. This shortcut removes useful friction, but it also hides an assumption that should be tested. Replace it with an observable check.
  • Assuming specificity proves research. This shortcut removes useful friction, but it also hides an assumption that should be tested. Replace it with an observable check.
  • Confusing a useful draft with a verified answer. This shortcut removes useful friction, but it also hides an assumption that should be tested. Replace it with an observable check.

Read

Build the habit

Before you close the tab, capture a reusable habit for Prediction is not knowledge inside Why AI makes mistakes: name the observable check, the evidence you would open, and the stop condition. Rehearse it once on a low-stakes example, then once on a higher-stakes variant. The habit succeeds when you can explain the check without reopening this lesson. Target outcome: Explain why plausible language generation does not guarantee factual accuracy.

Go deeper

Before you start

Why this matters

A travel assistant states that a small museum opens every Monday, even though the museum has always closed on Mondays.

In the wild

See how this idea shows up as a product and a company — then come back to the lesson. Skills transfer across vendors.

Check your understanding

Page assessment

Answer from memory. Completion is saved from this evidence, not from opening the next page.

Local focus for Prediction is not knowledge (Why AI makes mistakes): write the smallest test that would falsify a confident claim on this page, name the evidence you would open first, and note who must approve if the cost of being wrong is more than a redo. Keep the note under ten lines so you will actually reuse it.

1. Explain the page’s core distinction without using the word “smart.”
2. Which fact, source, permission, or test would most change your judgment in the opening scenario?
3. Name one low-consequence use where a light check is enough and one high-consequence use where independent review is required.
4. What should a responsible user do when the available evidence cannot support the requested conclusion?

All responses are required.