Shadow AI: What Organizations Found When They Looked

Most security teams assume AI use is broadly under control. They have an acceptable use policy, a sanctioned tool, often blocking at the network level. The activity underneath tells a different story, and very few teams have looked at it directly. This session shows what surfaces in the first 30 days when they do.

• Which channel carries two thirds of shadow AI data, and why most tools miss it
• The two problems that repeated across five real 30-day assessments
• Why source code stopped dominating shadow AI data
• What the room said about AI policies, and why having one proves nothing