Artificial intelligence is profoundly transforming the financial sector, bringing gains in efficiency, security and personalisation of services. However, it also brings new risks, particularly in terms of explainability, fairness and cybersecurity. The adoption of the European AI Act in 2024 introduces a horizontal regulatory framework that supplements existing sectoral regulations. As the supervisory authority for the financial sector, the ACPR must adapt its methods, strengthen its expertise and cooperate with public, private and academic players to ensure effective surveillance of AI systems. This article explores the legal, methodological and operational challenges of this evolving environment.