The Dutch Authority for the Financial Markets has published a set of 12 building blocks to help audit firms deploy advanced audit tools, including generative AI, in a responsible and controlled way. The message is that trust and safety must remain central and that AI should support, not replace, professional judgement in the audit process. AFM notes rapid growth in innovative audit tooling driven by new technologies, the phase-out of legacy audit packages, client digitalisation and private equity pressure for efficiency, which is changing how audits are performed. The building blocks focus on putting in place robust risk management and information security, ensuring the quality of data and analyses, and applying tooling consciously and in a controlled manner, with the “base layers” in order before adopting AI. For GenAI specifically, AFM sets three additional expectations: humans remain ultimately responsible for audit work and interpretation, AI outputs must be traceable or reproducible, and data security must be prioritised so AI does not become a black box in the audit.
Dutch Authority for the Financial Markets 2025-12-11
Dutch Authority for the Financial Markets publishes 12 building blocks for controlled use of GenAI audit tools
The Dutch Authority for the Financial Markets has released 12 building blocks to guide audit firms in responsibly integrating advanced audit tools, including generative AI. Emphasizing trust and safety, the guidelines stress that AI should enhance, not replace, professional judgment. Key expectations include maintaining human responsibility, ensuring AI output traceability, and prioritizing data security.