10/20/2026
The first major AI incident at most companies won't come from a rogue AI. It will come from an authorized agent, using valid credentials, accessing real data, and simply taking an action no one realized it was authorized to take.
At 2026, we showed why in the agentic era demands a fundamentally different approach to trust.
Why the target is always the data, and why your storage must be an active participant in your resilience strategy
How News Corp treats AI access the same way they treat a new employee with elevated privileges, and why they prove agent safety before production
Why Commvault says everyone has a backup, but the real question is whether you can trust it, how clean it is, and how fast you can get back
The takeaway: "Trust is not a prompt. It's architecture."
The organizations that move fastest won't be the ones that trust AI blindly. They'll be the ones that build infrastructure worthy of trust.
This is the session your security team needs to see: https://www.youtube.com/watch?v=dbTIeXJLpPE