Skip to content

FRM Part II · FRM Exam Part II · Advances in Artificial Intelligence: Implications for Capital Markets Activities

A firm's AI model for detecting market abuse is trained on historical alerts labeled by past analysts. A validator notes that past analysts systematically under-flagged a certain product type, and the model now under-flags it too. Furthermore, the firm plans to retrain monthly on the model's own flagged cases as new labels. What is the most significant consequence of this retraining plan?

The plan creates a feedback loop that entrenches and can amplify the original label bias. Because the model's own under-flagging determines what becomes training data, missed product types never appear as positives. More data from a biased process does not correct it, and validation remains necessary.

  1. AIt removes the original bias because more data is always corrective
  2. BIt creates a feedback loop that can entrench and amplify the existing label biasCorrect
  3. CIt guarantees improved explainability
  4. DIt eliminates the need for independent validation

Explanation

Biased labels produce a biased model; retraining on that model's own outputs means under-flagged cases never enter the data as positives, so the bias is reinforced. More data from the same biased process is not corrective, and retraining does not improve explainability or remove the need for validation.

Did you get it right without looking?

One question tells you little. A timed set on Advances in Artificial Intelligence: Implications for Capital Markets Activities shows your real accuracy, how long you take and where you lose marks.

Practise Advances in Artificial Intelligence: Implications for Capital Markets ActivitiesFree diagnostic mock

More Advances in Artificial Intelligence: Implications for Capital Markets Activities questions