The Stockholm International Peace Research Institute (SIPRI) was pleased to host a public webinar on the peace and security implications of AI agents, following the publication of a new SIPRI Essay on this topic.
Cutting-edge artificial intelligence (AI) models are increasingly agentic, meaning they are increasingly able to act on their own and work together with other AI systems. These ‘AI agents’ could bring big benefits, including speeding up scientific research, but they also carry serious risks.
AI agents, when deployed and interacting at scale, may behave in ways that are hard to predict and control. They may also be vulnerable to adversarial attacks and malicious uses. This has implications for AI governance as well as for international peace and security. The international policy community needs to recognize and respond to the risks that emerge from interacting AI agents in short order. Agentic AI is still in its infancy, but the window of opportunity for effectively ensuring that these systems are deployed in responsible ways may soon close.
The 60-minute webinar convened a panel of experts for a moderated discussion on the implications of AI agent interactions for international peace and security. Panellists discussed what AI agents are, how their interactions might create new pathways of risk, and what kinds of governance responses are needed.
Speakers
Alan Chan, Centre for the Governance of AI
Joel Z. Leibo, Google DeepMind
Sarah Shoker, Berkeley Risk and Security Lab
Moderator
Dr Vincent Boulanin, Director of the SIPRI Governance of AI Programme
Click here to register on Zoom.