What does meaningful accountability look like in this new landscape? And how can organisations equip developers to build trustworthy AI from the outset?
As AI systems grow more capable and autonomous, regulatory responses are evolving that place the burden of human oversight on users who deploy or work with AI. Yet, as agentic AI and multi-agent systems increasingly act, negotiate, and make decisions on our behalf, their actions move further from meaningful user control. AI developers make the foundational choices and recommendations that influence downstream risk. This event presents insights from organisations deploying agentic AI at scale to spark a transdisciplinary and multi-stakeholder conversation on what meaningful developer accountability looks like and how it can be integrated within organisations, regulatory regimes, and international governance frameworks.
Held alongside the UN Global Dialogue on AI Governance, this working lunch brings together leaders from industry, research, and governance to examine how developer accountability can become a more central pillar of AI governance. The discussion comes at a particularly timely moment, following the presentation of the UN Scientific Panel'sPreliminary Report, which highlighted how many of AI's most significant challenges are embedded in design choices, underscoring the need for governance frameworks that reach further upstream.
speakers
Sarah Mathews, Group Head Responsible AI, The Adecco Group
Daniel Dobos, Research Director, Swisscom
Brigitte Roy, Head of Partner Sales – DACH & France, Cognizant
Moderated by Gudela Grote, Co-Chair of the TASC Platform and Professor Emeritus of Work and Organizational Psychology at ETH Zurich
Hosted by Kitrhona Cerri,Executive Director, TASC Platform
Drawing on practical experience from organisations developing and deploying AI at scale, the conversation will explore how responsibility is distributed across today's increasingly complex AI supply chain. As organisations integrate foundation models, third-party tools, and agentic AI systems, the distinction between developers, deployers, and users is becoming increasingly blurred.
Find out more and join us by registering with the link above.