The Croatian Financial Services Supervisory Agency reported that its President of the Board, Ante Žigman, spoke at an international conference in Malta on operational resilience and said the use of artificial intelligence does not change firms' existing business responsibilities. In the panel discussion, he said senior management must be aware of AI-related risks, ensure those risks have been addressed, and manage AI as a new type of operational risk within firms' control frameworks. The discussion focused on operational risks linked to AI, including inaccurate audits, hallucinations, unauthorised actions, and effects on data, privacy and security. Žigman said the European Union Artificial Intelligence Act sets basic control rules and risk-based application areas, while autonomous systems' access to corporate and personal data should be subject to regular approval procedures under existing cyber risk management and General Data Protection Regulation requirements. He also said management should define procedures and responsibilities for operations, risk management and oversight during AI development and implementation, with enhanced monitoring of AI process outcomes based on each firm's measured risk and the system's classification under the EU act.