The Adoption of Artificial Intelligence as a Lever for Rationalizing Financial Decision-Making: An Analysis of Its Influence on Behavioral Biases, Risk Management, and Decision-Making Performance in the Moroccan Banking Sector

Résumé

In financial decision making, bank professionals' bounded rationality in making decisions and their confirmation, over-confidence and anchoring judgment lead them to make financial choices in banks. This research is designed to find out whether Artificial Intelligence (AI) can help with rationalizing the decisions and how it can be used for that. Multiple regression was used to analyze a survey of 200 professionals in the Moroccan commercial, participative and investment banks and asset management companies as well as 5,000 bootstrap resamples for estimating the parallel mediation model. The level of adoption of AI is positively correlated with the perceived bias control and risk management capability and the perceived decision rationality (p<0.001). The overall model explains 67.3% of the variance in the decision making performance. Decision rationality is the dominant channel (indirect effect = 0.332; 95% CI [0.238; 0.432]), followed by risk management (0.157) and bias control (0.053). The three mechanisms jointly account for 65.3% of the total effect of the adoption of AI, and a significant direct effect remains. But the debiasing effect of AI is the weakest link – it's essential to rationalist banking decisions by structuring the decision process and improve risk assessment. The results have implications in the field of AI governance in Moroccan banks and for the supervisory authorities.