A forthcoming artificial intelligence (AI) blueprint at the National Geospatial-Intelligence Agency (NGA) is designed to operationalize AI across every facet of the organization, from intelligence collection and analysis to budgeting and strategic risk assessment, according to an NGA official.
Speaking at the Defense One Tech Summit on June 16, Rear Adm. Michael Baker, associate director for operations at NGA, offered new details on the blueprint, which aims to serve as a roadmap for integrating AI across the agency.
“The foundation of that [blueprint], it’s really important to understand, is that it all starts with the concept of being a data agency,” Baker said.
According to Baker, NGA has evolved from an imagery agency to a geospatial intelligence agency and is now focused on turning “every single thing we do” into data through what he called a “virtuous data cycle.”
“So, if I take an image down and I’m an analyst looking at that, the first thing I’ve done is run computer vision through it,” Baker said. “The computer’s already helping me to know which images to look at … but as I put my human expertise into that, I’m pinpointing … what it means.”
Those analyst observations are then structured and fed back into AI systems, allowing NGA to process information more quickly and improve future analysis, Baker said.
The blueprint extends beyond mission operations to agency operations. Baker said NGA is using AI to support workforce management, performance reporting, financial management, and capital planning.
“The really important part for us right now is on the financial management and capital management side, because we’re applying business tools of AI to understand where are we spending our money? How is that tied back to that mission line? Where can we be more efficient?” he said.
Baker said the most interesting – and challenging – aspect of the blueprint is using AI to support executive-level planning and risk-based decision-making.
“How do you take in all of the insatiable requirements that the intelligence community gets demanded of? How do you align that to your task and your resources, understand the gap, and then turn that into a characterized risk understanding, so we can make risk-based decisions,” he said.
“That’s the AI blueprint,” he concluded, adding, “It’s a challenge for us, but it’s a really good opportunity.”