Skip to content

FRM Part II · FRM Exam Part II · The Financial Stability Implications of Artificial Intelligence

A bank's AI credit-scoring model shows a decline in accuracy over 12 months because customer behaviour and the economy have changed since training. Which action best addresses this specific problem?

Ongoing monitoring with drift detection and defined triggers for governed recalibration or retraining is the right response. Accuracy decays when customer behaviour and conditions change, so initial validation is not enough and unvalidated changes add model risk.

  1. AImplement ongoing performance monitoring with drift detection and triggers for recalibration or retraining under governance approvalCorrect
  2. BIncrease the model's complexity to capture more patterns without re-validation
  3. CStop monitoring because accuracy at deployment was acceptable
  4. DReplace the model with a vendor model without any validation

Explanation

The problem is model or data drift, so the remedy is continuous monitoring with defined triggers and controlled retraining or recalibration. Adding complexity without validation increases model risk. Initial accuracy does not guarantee future performance, and unvalidated replacement shifts risk.

Did you get it right without looking?

One question tells you little. A timed set on The Financial Stability Implications of Artificial Intelligence shows your real accuracy, how long you take and where you lose marks.

Practise The Financial Stability Implications of Artificial IntelligenceFree diagnostic mock

More The Financial Stability Implications of Artificial Intelligence questions