TechNet Cyber 2026

Sessions Speakers Exhibitors Products Press Releases Floor Plan

The Collaboration Layer Is the AI Risk No One Is Managing (Room Theater #1 - Booth 3302)

03 Jun 26
10:45 AM - 11:45 AM

Tracks: Artificial Intelligence, Theater 1, Theater Sessions

Across defense and government, AI agents are being deployed into mission-critical workflows at a pace that governance frameworks haven't kept up with. The tools driving this shift — agentic AI, MCP connectors, large language models integrated with operational data — are delivering real capability. They are also creating a risk that most programs haven't named yet, and it isn't in the security stack.

This session challenges the assumption that existing Zero Trust and identity frameworks are sufficient for AI-enabled environments. Agents don't live in your SIEM or your EDR. They live in your collaboration platform — connected to mission data repositories, NIPR/SIPR knowledge management systems, C2 and ISR feeds, and operational infrastructure. Those platforms were designed for human communication. They were not designed for autonomous agents making decisions across classification boundaries. A trained analyst knows instinctively what data can cross which lines. Agents don't. They use everything they can access, in every context where it might be useful, unless the architecture explicitly prevents it.

Through documented real-world incidents and a concrete architectural framework, this session examines what it actually takes to govern AI agents in environments where classification boundaries, need-to-know, and operational continuity are non-negotiable. Attendees will gain a clear model for sovereignty-first architecture — built on customer-defined deployment at every classification level, interoperable lock-in-free design with full CAC/PIV and PKI integration, and Zero Trust extended to agents as distinct governed principals with per-agent identity, task-scoped delegation, and human-in-the-loop controls.

The session draws on current research from NIST NCCoE, ISACA, and DoD Zero Trust implementation guidance, and examines the emerging regulatory framework under the NIST AI Agent Standards Initiative and OMB M-25-21. Attendees will leave with a practical framework for evaluating collaboration platforms against mission requirements — and three diagnostic questions to bring directly into their next platform evaluation or RFP.