Project Maven
A Pentagon program (begun 2017) using artificial intelligence to locate, identify, and prioritize military targets from sensor data (such as satellite imagery), reducing time and personnel required for target identification; demonstrated ability to identify enemy tanks via satellite and coordinate strike systems in exercises
As defined in Inside the Pentagon’s $179B plan to win the next war
Where it comes up · 2
Most important claims · 7 of 7Search all claims for “Project Maven” →
- Project Maven reduces the time and manpower required to achieve positive identification of targets, which is operationally critical in contested environments saturated with enemy air and ground equipment; currently a human remains in the kill chain and AI systems are not autonomously launching weapons, though this constraint should be monitored.Inside the Pentagon’s $179B plan to win the next war
- Project Maven, operating since 2017, has been used to locate and identify military targets using artificial intelligence, with a 2020 test exercise (Scarlet Dragon) demonstrating the system identifying enemy tanks from satellite imagery and tagging them for destruction by HIMARS systems.Inside the Pentagon’s $179B plan to win the next war
- The Pentagon is budgeting over $1.8 billion for AI and autonomy projects including Project Maven (AI-assisted targeting and decision tools), AI-enabled drone autonomy for scouting, jamming, and targeting, and predictive maintenance/logistics automation; Project Maven specifically demonstrated identifying enemy tanks via satellite imagery and coordinating High Mars strikes in a 2020 Scarlet Dragon exercise.Inside the Pentagon’s $179B plan to win the next war
- The Pentagon is budgeting over $1.8 billion for AI and autonomy projects in FY2026, including AI-assisted targeting and decision tools under Project Maven and J-ACTA, drone autonomy for scouting/jamming/targeting, and predictive maintenance and logistics automation.Inside the Pentagon’s $179B plan to win the next war
- Talent is now the most effective constraint on big tech companies: Facebook and Google can afford to upset the US government but cannot afford to upset their 100 best machine learning engineers, as shown by the Project Maven walkout.Investing Pre Company
- Project Maven reduces the time and manpower required to obtain positive identification of military targets, which is critical in environments where rapid target identification can mean the difference between life and death.Inside the Pentagon’s $179B plan to win the next war
- Currently, humans remain in the military kill chain in Project Maven systems, meaning humans still make the final decision to launch weapons, though the boundary between human decision-making and automated targeting is shifting and should be monitored.Inside the Pentagon’s $179B plan to win the next war