The International Organization of Securities Commissions published a consultation report on how artificial intelligence, including large language models and generative AI, is being used across capital markets and the associated risks for investor protection, market integrity and financial stability. The report is positioned as IOSCO’s first-phase factfinding to inform whether further IOSCO tools or guidance are needed. IOSCO’s analysis, based on member and industry surveys and stakeholder roundtables, finds expanding AI adoption in decision-support and front-office functions (including robo-advice, algorithmic trading, investment research and sentiment analysis), alongside growing use in surveillance and compliance (notably anti-money laundering and counter-terrorist financing) and internal productivity (automation, coding support, summarisation and chatbots). Risks are organised around four recurring themes: malicious use (cyberattacks, fraud, market manipulation and deepfakes), model and data vulnerabilities (explainability, bias, hallucinations, data quality and drift), concentration and third-party dependency (infrastructure, data aggregators and model providers), and human–AI interaction risks (accountability, skills gaps and automation bias). The report also describes evolving firm governance and risk-management practices and a spectrum of regulatory responses ranging from technology-neutral application of existing rules to AI-specific guidance and bespoke AI frameworks. Comments are requested by April 11, 2025. IOSCO states that a second phase of work will consider whether to develop additional tools, recommendations or considerations for members, alongside ongoing coordination with other international bodies such as the Financial Stability Board.

Original source View in tracker