ST. LOUIS — Artificial intelligence continues to loom large over the future of U.S. geospatial intelligence. Speaking at the GEOINT Symposium on May 18, National Geospatial-Intelligence Agency (NGA) Deputy Director Brett Markham said the agency is making strides with AI but emphasized that full-scale adoption is still a work in progress.
“While we’ve made some significant progress in the last 12 plus months, there’s more work to be done,” Markham said.
NGA is tasked with collecting, analyzing and distributing geospatial intelligence (GEOINT) for both national security and combat support missions. Its analysts process satellite imagery and other spatial data to monitor global activity — from tracking troop movements and locating military hardware to supporting humanitarian disaster response.
At the center of NGA’s AI strategy is Maven, a program originally launched in 2017 to bring computer vision into defense intelligence workflows. Under Maven, AI is used to sift through vast amounts of imagery and video data, identifying potential military targets like vehicles, aircraft, or infrastructure. These tools aim to augment analysts’ capabilities by automating initial detection tasks and flagging anomalies faster than humans can.
Markham insisted NGA seeks to accelerate the integration of AI across analytical and operational processes, and wants AI not just in object recognition, but embedded throughout its data collection and orchestration systems.
“It is imperative that we adopt a more aggressive integration strategy relative to artificial intelligence,” he said. “More AI-driven tools need to be assimilated in the agency’s collection orchestration processes in order to keep up with the demand and to be able to efficiently and accurately task our systems.”
That demand is growing. According to Markham, NGA’s workload is expanding in response to a mix of geopolitical drivers, including the Russia-Ukraine war, instability in the Middle East, China’s global activities, and broader military requirements from regional combatant commands. The agency also contributes GEOINT for domestic and international disaster response.
Building infrastructure
To keep pace, NGA is building out a broader AI infrastructure, including standardized training data (image labels), reusable models, and guidelines for “responsible AI usage.” These efforts aim to create a scalable and reliable foundation for deploying AI in operational settings, meeting both technical and ethical expectations.
Markham acknowledged that while advances in machine learning and computer vision have been significant, the agency still hasn’t unlocked the full value of its data.
“I don’t think we fully tapped into the capability that is needed to be able to interrogate, manage and manipulate the data that we have before us now,” he said. “I think we’re on the early stages of how AI integration and adoption is going to impact this profession, and I’m quite excited about what we have seen thus far.”
To engage with the AI industry, NGA is hosting several upcoming events. On June 9, it will hold an AI Industry Day at its Springfield, Virginia, headquarters. The agency will provide updates on AI development and identify technology gaps, said Markham. Companies are invited to submit white papers in advance; selected proposals may be invited to an upcoming AI Pitch Day.
Looking ahead, NGA will convene a GEOINT AI Summit in November, also in Springfield, to discuss the future of geospatial intelligence.
Markham noted that NGA’s workforce will need to adapt to this new era. “We must expand the acquisition, the integration and adoption of artificial intelligence across all of our mission areas,” he said. “This will enable us to scale, to analyze more images and locate targets of interest quicker.”
