ShorePoint’s Ian Lee Publishes New Analysis on AI Readiness in Classified Federal Environments
Article drawn from DOE’s NLIT 2026 Summit explores foundational data preparation challenges.
ShorePoint’s Director of Advanced Computing Solutions, Ian Lee, has published a new analysis in OrangeSlices examining the foundational challenges federal national laboratories must address before artificial intelligence can deliver on its promise in classified environments. The piece, drawn from a panel Lee participated in at the 2026 NLIT Summit, argues that the most pressing obstacle to deploying AI against classified archives is not architectural. It is the state of the underlying data itself.
“Federal agencies and national laboratories are under real pressure to move faster with AI, and that urgency is justified,” said Ryan McCullough, Chief Growth Officer at ShorePoint. “The organizations that will ultimately succeed are the ones willing to do the foundational work first. Getting the data ready is not the glamorous part of this effort, but it is the part that everything else depends on.”
The federal government holds decades of irreplaceable scientific and mission data that AI could make genuinely accessible for the first time. In the article, “Before You Can Ask the Question, You Have to Own the Data,” Lee lays out some challenging realities about data quality, governance, and classification guidance to be addressed before access to that valuable data may be realized at scale.
Lee’s analysis outlines a three-tier architecture discussed at the NLIT Summit, a leading event for IT and cybersecurity professionals affiliated with the U.S. Department of Energy national laboratories. Lee and the other panel experts outlined the architecture structure, where data storage, access governance, and AI reasoning are separated so that AI systems never directly access classified archives.
The approach reflects sophisticated thinking about how to govern and protect classified data at scale. However, Lee argues its promise depends entirely on foundational work that must come first, including accurate and machine-readable metadata across classified archives and authoritative classification guidance consistently interpreted across organizational units. Additionally, data rights frameworks must be resolved for any shared or multi-party AI infrastructure. Without those foundations in place, the governed knowledge layer that sits at the center of the architecture “cannot make reliable access decisions regardless of how sophisticated the access control technology is.”
“The question of which AI architecture to adopt is premature if the underlying data cannot support it,” writes Lee. He goes on to pose three questions to guide federal leaders’ next steps before investing in the AI infrastructure.
The full article is available now on OrangeSlices. Read the full piece here.
About ShorePoint
ShorePoint is an elite, fast-growing cybersecurity services firm dedicated exclusively to strengthening the cyber resilience of federal agencies and their missions. With deep expertise and a forward-looking approach, ShorePoint’s experts operate where tomorrow’s threats are already taking shape — from AI and high-performance computing security to supply chain assurance — helping customers stay ahead of an evolving threat landscape. ShorePoint is based in Herndon, VA. www.shorepointinc.com









