FRM Part II · FRM Exam Part II · Advances in Artificial Intelligence: Implications for Capital Markets Activities
A regulator is considering how to supervise AI use in capital markets when many firms rely on the same few foundation-model providers. Which concern most directly motivates supervisory attention to this arrangement?
The main concern is concentration: if many firms depend on a few model providers, errors, similar signals or outages can hit them simultaneously. This creates herding, correlated behaviour and single points of failure, which makes it a systemic and third-party risk issue for supervisors.
- AConcentration in a few providers could create correlated errors, herding and operational single points of failure across firmsCorrect
- BFoundation models are always less accurate than in-house models, which lowers firm profitability
- CProvider concentration reduces the volume of data available to firms, which raises their capital requirements automatically
- DUsing common providers makes it impossible for firms to hold any capital against operational risk
Explanation
Shared dependence on a few providers can cause many firms to act on similar outputs or fail together after an outage, raising systemic and third-party risk. The other options assert unsupported or false mechanical effects.
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 firm adds an L1 (lasso) penalty to a regression-based machine learning model with 200 candidate predictors. Compared with an unpenalized m…
- A trading desk deploys a deep neural network to generate short-term signals. Validators find that the model performs well in backtests but i…
- A trading firm uses a reinforcement-learning algorithm for execution. A risk manager worries that the algorithm could learn behaviour that r…
- A trading firm's AI surveillance model was trained on two calm years of order-book data. During a sudden volatility spike, alert patterns ch…
- A trading desk deploys a deep neural network to generate execution signals. The model performs well in testing, but the risk committee canno…
- A risk manager reviews an AI model that screens equity trades using alternative data (social media sentiment) supplied by a vendor. Which da…