FRM Part II · FRM Exam Part II · Advances in Artificial Intelligence: Implications for Capital Markets Activities
A risk manager reviews a generative AI tool used to summarize market research for portfolio managers. The tool sometimes states plausible but fabricated figures with confidence. Which control is most appropriate to address this specific weakness?
Require human review and verification of the tool's figures against source data before they inform decisions. The weakness is hallucination, confident fabrication of plausible numbers, which is best controlled by human-in-the-loop checks. Longer outputs, self-training or removing logs would worsen or hide the problem.
- AIncrease the tool's output length so more context is provided
- BRequire human review and verification of cited figures against source data before use in decisionsCorrect
- CRetrain the tool exclusively on its own prior summaries
- DRemove all logging so outputs cannot be second-guessed
Explanation
Fabricated but plausible outputs (hallucinations) are best mitigated by human-in-the-loop verification against authoritative sources. Training on its own outputs can amplify errors, longer outputs add more room for errors, and removing logging weakens governance.
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.
More Advances in Artificial Intelligence: Implications for Capital Markets Activities questions
- A bank uses a reinforcement learning agent to execute large equity orders, minimizing implementation shortfall. During back-testing, it perf…
- A bank's risk committee is assessing AI-related vulnerabilities. Which of the following would be the strongest example of an AI-related chan…
- A dealer uses generative AI to summarize research and draft trade ideas for traders. Which risk is most specific to this use of generative A…
- A risk manager is reviewing a bank's use of a complex deep-learning model to flag potentially manipulative trading. Regulators ask the firm …
- A risk manager notes that a generative AI tool used to summarise regulatory changes sometimes produces confident but fabricated citations. W…
- A broker-dealer replaces its keyword-based trade surveillance rules with a machine learning model that scores orders for potential spoofing.…