As artificial intelligence becomes increasingly embedded in financial services, AI governance and supervisory expectations are moving to the forefront of regulatory and industry agendas across Europe.
These themes were at the center of the roundtable “AI Governance & Safe AI in Europe”, held during SupTech Week 2025 and organised by the Cambridge SupTech Lab in collaboration with the World Economic Forum.
Moderated by Karsten Wenzlaff, Board Member of the European Digital Finance Association (EDFA), the session brought together regulators, academics, and industry practitioners to discuss how emerging AI regulation, technical standards, and organisational governance frameworks can support safe and scalable innovation in financial services.
From regulation to implementation
A recurring theme was the challenge of translating frameworks such as the EU Artificial Intelligence Act into practical governance models that align with existing financial regulation, model risk management, and data governance obligations. Panelists highlighted the importance of proportionality, lifecycle oversight, and avoiding parallel compliance structures for AI.
Measuring and managing AI risk
Jeremy Sulzbacher, Chair of EDFA’s AI Working Group, presented insights from EDFA’s work on AI risk taxonomy, outlining how different fintech and financial services use cases map to varying AI risk profiles. This approach aims to support clearer risk classification and more consistent supervisory dialogue.
Academic perspectives emphasized the need to move beyond abstract principles towards measurable and comparable AI risk indicators, enabling both firms and supervisors to assess effectiveness and compliance in practice.
Supervisory perspectives and sandboxes
Supervisory representatives discussed the role of principles-based supervision, regulatory sandboxes, and supervised experimentation environments in supporting innovation while maintaining trust and accountability. The use of synthetic data and controlled testing environments was highlighted as a promising avenue for reducing uncertainty in AI deployment.
Data governance as a foundation
Across all perspectives, there was strong alignment on one point: robust data governance remains the foundation of safe AI. Issues around data quality, documentation, monitoring, and human oversight were consistently identified as critical success factors for compliant AI adoption in financial services.
Watch the full discussion
▶ Watch the SupTech Week 2025 roundtable: “AI Governance & Safe AI in Europe”
EDFA will continue to support dialogue between regulators, industry, and academia to promote practical, proportionate, and innovation-friendly approaches to AI governance across Europe.