Last year, "Dark Entry" tore down myths about "secure" infrastructure. This year, we have seen hasty AI integration and subsequent weaponization dramatically widen the risk aperture. This vulnerability stems from the rigid, legacy code rotting underneath it all, making modern encryption merely table-stakes for data security.
If the United States is going to win the AI race economically and from a national security perspective, we must face a harsh reality: forcing advanced AI agents into ancient architectures creates massive security blind spots, shadow AI usage, and catastrophic data leaks. Worse, there simply isn't enough time to manually rewrite this software, even with the help of AI.
The baseline reality of your current codebase is critical:
Cryptographic Inventory: Hardcoded and invisible. You cannot migrate what you have never counted.
AI Integration: Lacks the structural modularity and governance APIs needed to stop autonomous agents from leaking credentials or sensitive data.
Legacy OT & Weapons Systems: The most difficult assets to modernize, the last to get funded, and the first an adversary will pry open.
In this session, we will explore the depths of this legacy technical debt, which makes it untenable to patch or scaffold your way into the AI era. Surviving this landscape requires treating modernization as a continuous lifecycle: discovering hardcoded dependencies, refactoring application logic for instant crypto-agility, and engineering policy-driven guardrails directly into your code to govern what AI can reach and export. Modernizing your codebase to enable efficient, secure AI operations the definitive baseline for tactical survival and dominance on the cyber-physical front.