Skip to content

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

A broker-dealer deploys a generative AI tool that drafts client investment recommendations. The compliance head wants a governance control that most directly addresses the risk that the tool produces plausible but factually incorrect statements (hallucinations) that reach clients. Which control is most appropriate?

Qualified human review and sign-off of AI-generated recommendations before release, with logged overrides, is the best control. Hallucinations are plausible but incorrect outputs, so a human-in-the-loop check catches errors before clients see them, whereas more data, a cheaper vendor or IT-only access do not.

  1. ARequiring qualified human review and sign-off of AI-generated recommendations before release, with logging of overridesCorrect
  2. BIncreasing the size of the model's training data set to include more historical client communications
  3. CReplacing the tool's vendor with a cheaper provider that offers faster response times
  4. DRestricting access to the tool to the firm's IT department only

Explanation

Hallucinations are outputs that look credible but are wrong, so a human-in-the-loop check with documented review directly catches them before clients are affected. More training data does not eliminate hallucination. Changing vendor on cost or speed, or limiting access to IT, does not address output accuracy.

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