Everyone’s talking about frontier models finding zero-days.
Field Effect CEO and Founder, Matt Holland, is more concerned about something already happening inside your network: AI adoption that is moving faster than anyone can govern it.
Across Field Effect’s customer base, 94.9% of organizations are now using AI, 88% have found shadow AI on their network, and 6.53 TB of data went to AI vendors in just 30 days.
The problem is that these tools are non-deterministic. They may not do the same thing twice, and what they do under the hood often doesn’t show up in the logs.
In this recording of his SecTor 2026 main stage talk, Matt shares real customer cases and a first-hand test of popular AI tools, then lays out a practical way to get visibility and control.
What you’ll learn
- How employee AI misuse, attackers hijacking your AI, and the AI tools themselves create three distinct risks
- What a simple “read this file” prompt triggers in popular AI coding tools
- A zero-trust approach to securing AI: see it, block it, go deeper, enforce it
