The Guardrails Beneath the Guardrails
10 September 2026. Washington DC.
Cyber Guardrails for the AI Era, hosted by Squadra Ventures, ARK, Carahsoft, and PwC alongside the Billington CyberSecurity Summit, brought together participants from government, industry, investment, and the startup community for an after-hours panel and fireside chat. The room included representatives from the White House and DoD, along with cybersecurity leaders and founders. Kendall Moore of Sicura, Logan Havern of Datalogz, and Miles Parry of MPCH joined the panel, covering security controls, data access, analytics sprawl, cryptographic infrastructure, and the pressures AI is placing on existing systems.
During a fireside conversation with Squadra managing partner Guy Filippelli, Sean Coyne, Senior Advisor to the Director of the Cybersecurity and Infrastructure Security Agency (CISA), made a comment that captured one of the evening’s central concerns.
“Technical debt is the bane of my existence.”
The line stood out because technical debt is often treated as an engineering concern, even when its effects reach much further. A team may know that a vulnerability needs to be fixed or that a dependency should be replaced, yet accumulated complexity can make the consequences of that change hard to predict. Documentation may be incomplete, ownership may have shifted, and systems may have grown connected in ways that few people understand. At that point, technical debt begins to limit how confidently an organization can respond.
AI increases the speed of that process. More people can now build useful software, integrations, automations, and internal tools without the structures that once surrounded a conventional software project. That can create real value while also allowing software to become operationally important before ownership, review, and maintenance protocols have been established.
That is also why technical debt is not just a technical problem. The debt appears in code, infrastructure, testing, and dependencies, while the conditions that produce it often come from culture. Teams respond to deadlines and incentives, as well as budgets and management priorities. Maintenance loses out to new work. Ownership becomes vague and documentation gets deferred until the software has become critical to the business, or until something goes wrong. Over time, those habits determine how much control an organization retains over the software it depends on.
That is where technical debt meets cybersecurity. Security depends on identifying risks, and it also depends on whether an organization understands its systems well enough to change them when needed. A patch, migration, or new control has limited value if the surrounding system cannot absorb the change without creating another problem. The risk grows when AI-assisted development or vibe-coded prototypes enter the operational environment without clear oversight. A sound AI governance regime connects the capabilities of AI with the practical demands of operating, securing, and maintaining the systems a business depends on.
In the AI era, organizations can create software faster than ever. Their ability to understand, govern, and maintain it may prove just as important as the speed at which they can build it. That is where technical debt management and cybersecurity begin to overlap, in the shared need to preserve control over systems as they grow more complex.