IOSCO publishes new supervisory toolkit on the use of AI in capital markets

IOSCO has published a new guidance document for supervisory authorities entitled the “Supervisory Toolkit for AI Use in Capital Markets” (FR/02/2026). The document forms part of a multi-stage approach and builds on two earlier publications: the report ‘The Use of AI and Machine Learning by Market Intermediaries and Asset Managers’ (FR06/2021) and the report ‘AI in Capital Markets: Use Cases, Risks, and Challenges’ (CR/01/2025).

While previous publications primarily described the application areas, risks, and challenges of artificial intelligence in the financial sector, the new toolkit takes a significantly more practice-oriented approach. It provides supervisory authorities with concrete tools for inspections and oversight activities, including structured questions for on-site examinations and a standalone questionnaire for direct use in supervisory proceedings.

Another key focus is the technological evolution of AI systems. For the first time, IOSCO explicitly addresses Generative AI (GenAI) and Agentic AI, thereby acknowledging the growing adoption and increasing use of these technologies in capital markets.

In addition, the toolkit examines the impact of AI on cybersecurity. IOSCO notes that AI-enabled attacks can accelerate the development of threat scenarios and expects market participants to continuously assess vulnerabilities and regularly review existing security measures.

The Supervisory Toolkit does not replace existing national requirements. Rather, it is intended to complement existing regulatory and supervisory frameworks and to assist supervisory authorities in evaluating the use of AI in regulated firms. At the same time, the publication provides financial market participants with insight into the topics and questions that are likely to become an increasing focus of future supervisory reviews and examinations.