IAPP Europe Congress 2026
Privacy | AI governance | Cybersecurity law
BRUSSELS
16-19 November
GPA: Accountability by Design in AI Systems: Roles, Evidence and Recourse
Thursday, 19 Nov.
13:15 - 14:15 CET
Advanced level
Part of the Global Privacy Assembly Open Programme.
When regulatory investigations and civil exposure reach AI-enabled systems, formal role labels are the starting point, not the conclusion. What regulators and counterparties examine is who controlled what, when and whether the documentation exists to prove it. The operational mapping between GDPR accountability obligations and AI Act role allocation remains insufficiently specified for multi-party AI deployments and that ambiguity becomes a liability when systems are modified, responsibilities shift and recourse disputes emerge. This challenge is most acute in agentic AI systems, where control is distributed and accountability chains are harder to document by design.
This session builds the case for accountability by design: a governance architecture in which operational control is documented in real time, not reconstructed post-incident. Three recurring failure modes structure the dialogue: role drift following system modification and autonomous behavior, evidentiary asymmetry across providers and deployers, and the gap between contractual risk allocation and demonstrable oversight. They are examined through structured conversation with senior practitioners and regulatory perspectives on agentic AI governance. Attendees leave with a framework they can apply: governance design principles, evidentiary baselines and escalation architecture for high-impact and agentic AI processing.
What you will learn:
- How to close the gap between formal GDPR and AI Act role labels and actual operational control before a regulator or counterparty does it for you.
- The evidentiary baseline your company needs under concurrent regulatory and civil scrutiny and the governance checkpoints that produce it before a dispute arises.
- How to embed accountability checkpoints into your AI change management, including in agentic systems where change may be autonomously generated, so governance questions surface at the right moment.
Featured in this session

Davide Borelli
Counsel, Data and Technology
Freshfields

Margo Steiner
VP Global Privacy, Data and AI Regulation
adidas AG

Wouter-Bas van der Vegt
AIGP, CIPP/E, CIPM, FIP
Global Data Protection Officer; Director, Global Legal
Randstad