Current governance and risk management frameworks are not designed to address the system-level risks that can emerge when AI agents interact with one another at scale. Mitigating these risks will require coordinated action across the AI lifecycle. What interventions are needed, and which actors should be responsible for developing, implementing, and overseeing them?

Together with the International Association for Safe & Ethical AI (IASEAI), CAIF is hosting an online seminar that will present a blueprint for multi-agent AI governance.

The session will begin with an overview of the current state of multi-agent AI and the governance challenges it presents, followed by a panel discussion with experts in the field, including Patricia Paskov (Director of Standards at AVERI), Nada Madkour (Director of the AI Security Initiative at CLTC Berkeley), Sebastian Hallensleben (Chair at CEN-CENELEC JTC 21 and OECD), and Sam Boger (Research Fellow, AI Agents at Singapore AI Safety Hub).

We'll explore five key actions to help address multi-agent risks from advanced AI (Evaluate, Identify, Monitor, Report, and Incentivise), how they can be operationalised in practice, which actors are best placed to carry them out, and how responsibilities can be coordinated across governments, industry, civil society, and the research community.

This seminar builds on a workshop held during IASEAI ’26 by the Cooperative AI Foundation and the Brookings Institution, bringing together 40 experts across governance, policy, ethics, law, and technical research. For background, read the workshop memo: Establishing Foundational Principles and Thresholds for Multi-Agent AI Governance.

Subscribe to our Google Calendar to stay up-to-date on all upcoming events.