The Dutch Authority for the Financial Markets has published a sector report, AI in Capital Markets: Balancing Innovation and Integrity, on how AI is changing trading in capital markets and can shape market behaviour in ways investors may not readily detect. It highlights that, alongside efficiency gains, AI can amplify vulnerabilities and create new integrity risks, and it frames human involvement and clear responsibility as central to safe deployment. The report points to self-learning trading systems potentially generating price patterns that resemble coordinated behaviour without any agreement between parties, particularly where models draw on similar data and pursue similar objectives, allowing erroneous or defensive responses to spread rapidly without adequate controls. It also identifies information as a new vulnerability because models increasingly ingest news, sentiment and social media, shifting manipulation risk from orders and markets to the information feeding algorithms, where a strategically placed false report could theoretically trigger widespread automated reactions. The AFM also flags the rise of multi-agent technology, which may improve speed and efficiency but requires careful system design, risk management, defined boundaries and oversight of potential interventions, supported by transparent choices and identifiable human responsibility. To support these developments, the AFM will further strengthen its supervision and internal expertise, translate emerging risks into supervisory measures and consider whether additional interpretations or changes to laws and regulations may be needed. It also notes that rules and expectations for algorithmic trading and the use of AI have recently been tightened and that engagement will continue with the sector and with national and international supervisory counterparts.

Original source View in tracker