Journal of
Italian
Banking
Association

year 100
 

Archive » November 2024 » Artificial Intelligence, risk management and the financial sector: the Safe model to assess the risks of Ai

Artificial Intelligence, risk management and the financial sector: the Safe model to assess the risks of Ai

Paolo Giudici
November 2024 - n. 11
Keywords: Ai risk management, Accuracy, Explainability, Fairness, Sustainability
Jel codes: O30, O33, G32

The growth of Artificial Intelligence applications requires to develop risk management models that can balance opportunities with risks. We contribute to the development of Ai risk models presenting a set of integrated statistical metrics that can measure the Sustainability, Accuracy, Fairness and Explainability of any Ai application, in line with the requests of the European Ai Act. The proposed metrics are consistent with each other, as they are all derived from a common underlying statistical methodology: the Lorenz Curve and the related Gini index. They are very general and can be applied to any machine learning method, regardless of the underlying data and model. Their empirical validity is assessed in this paper by means of their practical application to a set of use cases. The application reveals that the proposed metrics are more interpretable and more consistent with the expectations, with respect to the currently used assessment metrics such as Mean Squared Error, Area Under the Curve, Shapley values and Fairness parity measures.

Interested in this paper?
Buy the issue