Banks split on human oversight of AI models

Risk USA: Most firms supervise their models, but one expert says they can be trusted to make decisions

Should machine learning models be allowed to make risk decisions?

Financial firms have differing views on whether machine learning models should be allowed to operate without human supervision.

Stephen Hsu, head of model risk management for Pacific Western Bank, said he would not want an unsupervised model to make automated decisions – and neither would regulators.

“You let the machine make the decision, probably the first question would be from the regulators