SESKStuff Everyone Should Know
SESK: Stuff Everyone Should Know · sesk.info/ai-unrepresentative-sample-bias

How can an unrepresentative training sample bias an AI system?

It can create systematic prediction errors for underrepresented cases.

Start playing free

  1. It creates systematic errors ✔
  2. It guarantees equal error rates
  3. It only slows training

Why: A model learns patterns from the examples it receives. If those examples poorly represent deployment, its errors can differ across people, places, or situations.

Huh, didn't know that: More data does not fix sampling bias if the added data has the same gaps.

The dad joke
The dataset skipped class, then wondered why it missed the lesson.
Like this card?
Play the game

References

Answer: It creates systematic errors. A model learns patterns from the examples it receives. If those examples poorly represent deployment, its errors can differ across people, places, or situations.

  1. NIST AI Risk Management Framework 1.0 nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf
    Computational and statistical biases can be present in AI datasets and algorithmic processes.
    Checked 2026-10-10.

Think this answer is wrong? Tap "Challenge this answer" on the card in the game and tell us why.

Test yourself in the game

Artificial IntelligenceWhat happens when you raise a language model's temperature setting?Artificial IntelligenceHow do AI, machine learning and deep learning fit together?Artificial IntelligenceWhat did Alan Turing call his 1950 test of machine intelligence?

That's stuff everyone should know.