Skip to content

ACCA Applied Knowledge · Business and Technology · Financial systems and technology

A bank uses a machine learning model to approve loans. Regulators complain that staff cannot explain why the model rejected particular applicants, and that rejection rates are higher for one group because of patterns in the historical data used to train it. Which pair of issues does this describe?

The scenario describes lack of transparency, often called the black box problem, and algorithmic bias. Staff cannot explain the model's decisions, and historical training data has produced unfair outcomes for one group. Downtime, licensing costs, phishing and denial-of-service attacks are not what is described.

  1. ALack of transparency (the 'black box' problem) and algorithmic biasCorrect
  2. BData redundancy and system downtime
  3. CScalability limits and software licensing costs
  4. DPhishing and denial-of-service attacks

Explanation

Being unable to explain model decisions is the opacity or 'black box' problem. Higher rejection of one group due to skewed training data is algorithmic bias. The other options relate to operational or security matters unrelated to the facts described.

Did you get it right without looking?

One question tells you little. A timed set on Financial systems and technology shows your real accuracy, how long you take and where you lose marks.

Practise Financial systems and technologyFree diagnostic mock

More Financial systems and technology questions