Skip to content

ACCA Strategic Professional · Strategic Business Leader · Enabling success: disruptive technologies

Halden Insurance deploys a machine learning model that automatically declines claims it scores as likely fraudulent, with no human review. Customers cannot see why claims were declined, and regulators are asking questions. Which governance response is MOST appropriate to address the underlying weakness?

The best response is to add human oversight of declined claims and require explainability and clear accountability for model outputs. The problem is opaque automated decision-making, for which the board stays responsible. Better accuracy, outsourcing or abandoning fraud screening would not resolve the transparency and accountability gap regulators are questioning.

  1. AIntroduce human oversight of declined claims and require explainability and accountability for model decisionsCorrect
  2. BIncrease the volume of training data to improve the model's accuracy score
  3. COutsource the model to a third-party supplier to transfer accountability
  4. DRemove all fraud screening to avoid regulatory scrutiny

Explanation

The weakness is a lack of transparency, explainability and accountability for automated decisions, which the board remains responsible for. Human oversight and explainability address this directly. More data does not make decisions explainable, outsourcing does not transfer accountability, and removing screening raises fraud risk.

Did you get it right without looking?

One question tells you little. A timed set on Enabling success: disruptive technologies shows your real accuracy, how long you take and where you lose marks.

Practise Enabling success: disruptive technologiesFree diagnostic mock

More Enabling success: disruptive technologies questions