Encore Presentation | How can you prevent AI agents from runaway behavior or privilege abuse? A permissions design framework for preventing AI incidents, learned from real cases
As companies rapidly expand their use of AI agents that can autonomously execute work, the security risks around those agents are becoming more visible.
Incidents caused by gaps in permission settings and unintended AI behavior are increasing, making "how much authority should we give to AI agents?" a critical question for many organizations.
In this webinar, we will review real incident cases and explain how to design AI agent permissions so organizations can continue using agents safely.
We will introduce a practical framework for balancing safety and usefulness, without locking down AI so tightly that it can no longer deliver value.
Participants will leave with security checkpoints they can use to review their own environments from the next day. This is a practical 30-minute session built for teams working with AI agents.
Please join us. This session repeats the same content from the previous month.
Webinar Overview
- Title: "AI Agent Security Guide"
- Date and time: Thursday, May 21, 2026, 13:00-13:30
- Format: Online via Zoom
- Fee: Free, with advance registration required
- Host: Air Company
- In cooperation with: QueryPie AI G.K.
Please register with your company email address.
Air Company may decline webinar participation at its discretion. Thank you for your understanding.
Register now