Skip to content

CS Professional · Artificial Intelligence, Data Analytics and Cyber Security - Laws and Practice · Artificial Intelligence - Introduction and Basics

An e-commerce company's fraud-detection model was trained on 2021 transaction patterns and performed well. By 2025 its accuracy has fallen sharply because customer behaviour and fraud tactics changed, though the code is unchanged. What is the best diagnosis and response?

The best diagnosis is data or concept drift, where real-world patterns change after deployment and a static model degrades. The response is continuous performance monitoring and retraining on recent data, not reducing the test set, deleting history or adjusting creativity settings.

  1. AData or concept drift; monitor performance and retrain on recent dataCorrect
  2. BOverfitting; reduce the size of the test set
  3. CAlgorithmic bias; delete all historical records permanently
  4. DHallucination; increase the model's creativity setting

Explanation

When the relationship between inputs and outcomes shifts after deployment, a static model degrades; this is drift. The remedy is ongoing monitoring and retraining or recalibration with current data. Shrinking the test set does not fix real-world change, deleting all history removes needed learning data, and hallucination relates to generative outputs.

Did you get it right without looking?

One question tells you little. A timed set on Artificial Intelligence - Introduction and Basics shows your real accuracy, how long you take and where you lose marks.

Practise Artificial Intelligence - Introduction and BasicsFree diagnostic mock

More Artificial Intelligence - Introduction and Basics questions