NGA Launches GEOINT-specific Artificial Intelligence Model Accreditation Pilot

SPRINGFIELD, Virginia - The National Geospatial-Intelligence Agency announces the launch of an accreditation pilot for GEOINT artificial intelligence models for the National System for Geospatial Intelligence: the Accreditation of GEOINT AI Models (AGAIM), pronounced A-GAME. Model accreditation evaluates the methodology and robustness of a program's model development and test procedures.

AGAIM will expand the responsible use of GEOINT AI models and posture NGA and the GEOINT enterprise to better support the warfighter and create new intelligence insights. Accreditation provides a standardized evaluation framework, implements risk management, promotes a responsible AI culture, enhances AI trustworthiness, accelerates AI adoption and interoperability, recognizes high-quality AI and identifies areas for improvement.

Additionally, in line with Department of Defense guidance on ethical AI, NGA established and launched a GEOINT Responsible AI Training (G - R - E - A - T) for all coders and users of GEOINT AI. NGA held pilot classes in April and May; the plan is for training to eventually be broadly available to anyone in the NSG coding or using GEOINT AI capabilities.

GREAT is tailored to developer- or user-specific challenges across the AI lifecycle, and everyone taking the certification will be asked to sign a final pledge to develop or use AI responsibly.

NGA has been using computer vision and machine learning for decades to manage the huge influx of data from satellite imagery. NGA's analysts have made significant advances in structured observation management, auto reporting and modeling. This integrated data approach enables more automated reporting, visualization, and advanced analytics, enhancing the speed and precision of GEOINT. Warfighters can directly access NGA computer vision and AI capabilities through Analytic Services Production Environment for the NSG (known as ASPEN) and Maven, and we've ensured NGA GEOINT AI models can run in other machine learning platforms.

# # #