New effort advances real-time model health monitoring and explainable AI in support of trusted autonomy for the Department of the Air Force
ResilienX, Inc. announced that it has been selected by AFWERX for a Small Business Innovation Research (SBIR) Phase I contract focused on explainable AI agents for real-time model health monitoring in uncrewed aircraft system (UAS) ecosystems to address a pressing challenge in the Department of the Air Force (DAF). The effort is sponsored by AFWERX, a component of the Air Force Research Laboratory (AFRL), through the Open Topic SBIR/STTR program. The award was made on June 30, 2026.
Modern autonomous systems increasingly rely on machine learning models whose behavior can change once they leave the lab and meet real-world conditions. As sensors, environments, and mission inputs shift in ways a model never saw during testing, that model can begin producing unreliable results with no outward sign that anything is wrong, leaving operators unable to tell a sound AI decision from a failing one. Through the Phase I effort, ResilienX will demonstrate how its FRAIHMWORK platform continuously tracks the health of deployed models, watching for early signs that a model’s performance is slipping and alerting operators before a quiet failure can affect a mission. When the platform flags a questionable result, it also delivers a clear, plain-language explanation of the reasoning behind it, so the people who depend on the AI understand what it is doing and can act with justified confidence rather than blind trust.
“Operators are being asked to trust AI they cannot see inside,” said Matt Synborski, Chief Technology Officer of ResilienX. “This effort puts real-time model health and plain-language explanations directly in front of the people who depend on these systems, so the Air Force can field autonomy with confidence rather than doubt.”
During Phase I, ResilienX will validate this approach with Department of the Air Force stakeholders and lay the groundwork for broader model-assurance applications across defense and government missions where reliable, understandable AI is essential.
The views expressed are those of the author and do not necessarily reflect the official policy or position of the Department of the Air Force, the Department of War, or the U.S. government.
