Overview
Drawing on Partnership on AI’s work on real-time monitoring, the Frontier Model Forum’s work on safety thresholds, and PAI’s forthcoming research on global AI assurance, the panel will focus on what’s needed next to support safe, trusted, and scalable AI deployment.
Key Discussion Topics
- Why governing AI agents can’t stop at pre-deployment testing
- What real-time monitoring looks like in practice and where it’s already being tested
- How threshold-based safety frameworks can enable faster, more credible oversight
- What it takes to turn frontier safety research into tools governments and developers can actually use
- Where today’s AI assurance ecosystem is falling short and how to close gaps across global contexts, including the Global South
Date & Time
February 20, 2026
5:30–6:25 PM IST
Location
Room 6, Bharat Mandapam