A loan officer currently makes approval decisions using judgment alone. Leadership wants to introduce an ML model that scores each application and shows the officer a risk estimate alongside the application, while the officer retains final approval authority. Which value does this design primarily deliver?
- A. Decision support that augments human judgment with data-driven insight Correct
- B. Full automation that removes the need for human review
- C. Elimination of all bias from the lending process
- D. A guarantee of regulatory compliance
Why A is correct
Surfacing a model-generated risk score to a human decision-maker who keeps final authority is the definition of decision support: the model augments human judgment with data-driven insight rather than replacing it.
Why the others are incorrect
Full automation would mean the model decides without human review — explicitly not the case here, since the officer retains approval authority. ML can reduce some inconsistency but doesn't eliminate all bias, because models can inherit bias from training data. And using an ML model doesn't by itself guarantee regulatory compliance — that depends on the overall lending process and its governance.