Artificial Intelligence & Machine Learning
,
Black Hat
,
Events

CAS Strategies CEO Camille Stewart Gloster on Authority and Human Oversight

Michael Novinson (MichaelNovinson) •
August 11, 2026    

Camille Stewart Gloster, CEO, CAS Strategies, and former deputy national cyber director, The White House

Autonomous artificial intelligence agents introduce a new class of insider risk because they can hold identities, credentials and authority while acting across enterprise systems, said Camille Stewart Gloster, CEO of CAS Strategies and former deputy national cyber director at the White House.

See Also: AI Phishing Now Frighteningly Normal, Hard to Detect

Traditional governance models focus on permissions and technical capabilities. Agentic AI requires organizations to define how much authority an agent can exercise, where that authority ends and which actions require human approval, Gloster said. Organizations also need runtime enforcement because policies alone can’t control agents pursuing objectives through unexpected methods.

“The goal is not to set a bunch of boundaries and hope they hold. You need to be able to enforce them at the point of action, and then the last, which ends up being the hardest for most organization, is continuous learning,” she said.

In this video interview with ISMG at Black Hat USA 2026, Gloster also discussed:

How agent inventories and critical-path controls preserve human oversight;
Why chief executives should maintain ownership of cross-functional AI governance;
Managing the trade-off between moving fast on AI and maintaining meaningful control over it.

Gloster advises governments, companies and nonprofits on AI, cybersecurity and digital governance risks as CEO of CAS Strategies. She previously served as the White House’s first deputy national cyber director for technology and ecosystem security, leading initiatives on AI security, quantum readiness and workforce resilience.