Duncan McCarthy has been with the National Geospatial-Intelligence Agency (NGA) for 32 years, where he has led Research and Development programs for the past 25. He holds a master’s degree in Information Science from George Mason University and has long focused on applying advanced technologies to critical national security missions.

In 2022, Mr. McCarthy was involved in research that revealed key limitations in artificial intelligence technologies for geospatial applications. Motivated by these findings, he launched a new initiative in 2024 to develop a geospatial generative AI foundation model designed to store and transact the types of information most vital to NGA’s mission.

His current work addresses the scientific challenges of applying large multimodal generative AI models—such as ChatGPT, Claude, Grok, Gemini, LLaMA/LLaVA, and Perplexity—to the domain of geospatial intelligence. These models, while proficient in verbal communication and increasingly capable with image data, face unique obstacles when working with satellite imagery and geospatial patterns. Mr. McCarthy and his team are developing new methods and benchmarks to measure and improve generative AI’s ability to recognize and process geospatial data effectively, paving the way for more reliable and scalable solutions in future intelligence systems.