The US National Geospatial-Intelligence Agency (NGA) has proposed a potential $708 million contract to advance the training of its AI-powered computer vision systems through data labeling.
Set to span up to seven years, the initiative represents the largest-ever investment by the NGA in bolstering its data labeling capabilities.
The NGA uses AI-powered computer vision models to analyze large amounts of satellite imagery and geospatial data from multiple sources, including commercial satellites.
Enhancing data labeling methods will allow these AI systems to more efficiently tag and categorize raw data, improving accuracy in identifying military targets and unusual activities.
The aim is to accurately spot military targets or unusual activities from satellite images and other geospatial data to advance national security, military operations, and disaster response.
The NGA seeks this improvement for multiple directorates and programs.
The agency said the contract will also provide training for the Maven Program, which was launched in 2017 to accelerate the adoption of AI into US military operations.