AI And The Future Of Supervision: How Artificial Intelligence Will Impact Financial Supervisory Authorities | Nicolas Vasse, Richard Gigax, Thanh Tran Tien & Sipke Hiemstra

Published (in revised form) on 13th October 2024 in Journal of Financial Supervisors Academy Volume I.

Artificial Intelligence is set to fundamentally reshape how financial markets are regulated and overseen, and supervisory authorities cannot afford to be passive bystanders. That is the central argument of an article by Nicolas Vasse (now Executive Director at AMLA), Richard Gigax (Regulatory Compliance Consultant), Thanh Tran Tien (Product Manager at Banque de France), and Sipke Hiemstra (Senior Principal Consultant at Promontory Financial Group), which offers a comprehensive examination of how AI - and Generative AI in particular - is transforming financial supervision.

The article traces AI's journey from neural network research in the 1950s to today's Large Language Models, identifying the arrival of tools like ChatGPT as a genuine turning point. Unlike previous technologies that took years to reach mainstream adoption, Generative AI landed in public hands almost immediately, with serious implications for regulated industries. The authors are clear-eyed about the risks this introduces: Generative AI is probabilistic and often opaque, meaning its outputs can be biased or incorrect even when they appear authoritative, a challenge that sits at the heart of any responsible deployment in financial services.

For supervisory authorities, the article identifies three distinct pressure points. Regulators must update and extend legislative frameworks; the EU's AI Act, with its risk-based approach, provides a baseline, but the authors expect sector-specific rules to follow. Supervisors must also ensure that financial institutions integrating AI into their operations do so without compromising compliance or introducing unmanaged risks, with Model Risk Management emerging as a key area of focus. And crucially, supervisors have a significant opportunity to deploy AI within their own work - from real-time transaction monitoring and market abuse detection to the efficient analysis of large document volumes during inspections. Examples already exist: France's ACPR has developed the LUCIA transaction analysis tool, while the ECB uses Heimdall to assess fit and proper questionnaires.

The authors' concluding call to action is pointed. Supervisory bodies are urged to build dedicated AI hubs, attract specialist talent, and collaborate actively with the private sector. Regulators who wait to be disrupted will find themselves outpaced. Those who lead the change will be far better placed to protect the stability and integrity of financial markets in an AI-driven world.

This summary is the third in a weekly series curated by Business News Malta, showcasing articles from the Journal of Financial Supervisors Academy (JFSA) Volume I, published September 2025.