Grady Summers, CEO of Netwrix, argues that widespread AI adoption is stress-testing enterprise security foundations, revealing a dangerous structural vulnerability: the separation between identity security and data security, according to Forbes.
Drawing on a survey of 2,150 IT professionals, Summers notes that while 60% of organizations now leverage AI tools, adoption often occurs outside formal governance. He warns that AI, operating at machine speed, exposes "identity debt" or accounts that authenticate perfectly but retain excessive permissions accumulated over time. Traditional "identity-first" models fail because AI doesn't simply access data; it consumes, transforms, and reshapes it. When an AI agent has overprivileged access, "the breach isn't a single event; it's a continuous leak."
Summers calls for "continuous contextual authorization" bridging identity and data, requiring three technical shifts: mapping the explosion of nonhuman identities, adopting zero-trust data that verifies transactions not just users, and auditing the AI-data loop where newly generated content inherits existing security policies. Organizationally, he urges shared accountability across security, data, legal, and AI stakeholders, policy definition before deployment, and investment in AI literacy.
