What is the top security measure for enterprise AI deployment?

Ric Smith

President of Products and Technology · Okta

Deterministic fine-grained access

Ric Smith says intent-based authorization starts with deterministic fine-grained access, which then narrows by intent, and says this is what his team is currently working on.

Sam Curry

Chief Information Security Officer, SVP · Zscaler

Reduce attack surface

Sam Curry says the focus should be on reducing the attack surface, so attackers have fewer options, rather than just applying more AI to defense and offense.

The answer starts at 5:29.

Mandy Andress

CISO · Elastic

Strong governance

Mandy Andress says the key is understanding and implementing strong governance, including guardrails, access controls, and intent monitoring, to keep agents from going rogue or accessing systems they shouldn't.

The answer starts at 6:05.

Jacob Thomas

Manager of Information Security · Texas Children'S Hospital

Good governance model

Jacob Thomas says a good governance model is the key security measure, because mitigating and compensating controls planned correctly can reduce the attack surface and a threat actor's ability to act.

JA

James J Azar

Chief Information Security Officer · Artera Services

Least privilege

James J Azar says least privilege should be a foundational architectural requirement for AI before deployment reaches production, arguing AI is becoming another privileged enterprise user that needs the same governance standards as human administrators.