Federal Emergency Management Agency
Geospatial Damage Assessments from aerial imagery
FEMA retired its AI system for assessing disaster damage
Problem
FEMA needed to assess structural damage after natural disasters and share useful information with recovery partners.
Approach
Its Geospatial Damage Assessments system used machine learning and computer vision to analyze aerial imagery. Officials told GAO that missing data rights prevented them from sharing model outputs with federal and state partners. DHS reported in February 2026 that FEMA had retired the system.
Results
No published figures
Takeaway
Data rights constrained the sharing of assessment outputs. GAO did not establish that this constraint caused retirement or disclose when operational use began.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer visionPredictive analytics and ML
- Adoption
- Vendor solution
- Evidence
- Regulator, audit or court
Sources
Source summaries
GAO described FEMA's aerial-image damage assessment system and reported its retirement. It separately documented officials' concerns about contractual data rights that prevented sharing model outputs with partners.
Full details are in the original sources.