Skip to content

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

A treasury officer at a mid-sized bank receives a video call that appears to show the CFO instructing an urgent large wire transfer. The voice and face are later found to be AI-generated. Which control would most directly address this type of AI-enabled fraud?

Out-of-band verification combined with dual authorisation of high-value payments is the most direct control. Deepfakes defeat voice and video recognition, so confirming through an independent channel and requiring a second approver stops a single forged instruction. VaR, credit scoring data and capital buffers do not address payment fraud.

  1. ARequiring out-of-band verification and dual authorisation for high-value payment instructionsCorrect
  2. BIncreasing the bank's VaR confidence level from 95% to 99%
  3. CUsing a larger training dataset for the bank's credit scoring model
  4. DRaising the bank's countercyclical capital buffer

Explanation

Deepfake impersonation attacks exploit trust in audio and video identity. Out-of-band verification and dual authorisation break the single-channel dependency and are direct process controls. The other options address market risk, model performance or macroprudential capital, which do not mitigate payment fraud.

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.

More The Financial Stability Implications of Artificial Intelligence questions