Skip to content

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

A risk manager reviews an AI credit-scoring model trained on 2015-2021 data, a period of low defaults and low interest rates. Which governance action best addresses the risk that the model's performance will degrade when conditions change?

The best action is ongoing monitoring with out-of-sample and stress testing and predefined recalibration triggers. A model trained in a benign period may fail when conditions shift, so continuous validation detects drift, whereas higher complexity or removing human oversight would increase risk.

  1. ARely on the original in-sample accuracy, since it was high at development
  2. BIncrease model complexity to reduce training error further
  3. CImplement ongoing performance monitoring with out-of-sample and stressed-scenario testing and defined triggers for recalibrationCorrect
  4. DRemove human oversight so the model adapts without delay

Explanation

Training data drawn from a benign period creates data drift and regime risk. Continuous monitoring, out-of-sample and stress testing, and recalibration triggers address this. More complexity usually worsens overfitting, and removing oversight reduces control.

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