While the House recommended cutting $50 million from the DoD’s Project Maven in fiscal 2021, House and Senate conferees on the fiscal 2021 National Defense Authorization Act, H.R. 6395, agreed to the Senate’s recommendation to fund the Pentagon’s $250 million request for the program–renamed as the Algorithmic Warfare Cross Functional Teams Software Pilot Program.
Congress funded Project Maven at $221 million in fiscal 2020.
The future years defense program (FYDP) blueprint calls for outyear funding of $252 million in fiscal 2022 for Project Maven, $120 million in fiscal 2023, $121 million in fiscal 2024, and $122 million in fiscal 2025. The decreases would presumably come as the military services begin fielding the artificial intelligence capabilities embedded in Project Maven.
The fiscal 2021 NDAA conference bill also negates the House’s recommendation to increase the Air Force $121.5 million research and development request for Airborne Reconnaissance Systems by $18 million–$10 million extra for the Sierra Nevada [SNC] Gorgon Stare Wide Area Motion Imagery (GS WAMI) surveillance system and $8 million more for sensor open systems architecture.
Conferees on the fiscal 2021 NDAA ended up acceding to an Air Force request to transfer the House recommended $18 million increase to research and development for the Lockheed Martin [LMT] U-2 Dragon Lady.
The House Armed Services Committee had wanted the $10 million Gorgon Stare increase to modernize the system. HASC said that the Air Force lacked a sufficient number of Gorgon Stare systems to satisfy combatant commander demands outside of U.S. Central Command, where Gorgon Stare “supports multiple daily orbits.”
“The committee is concerned that, despite daily operational tasking, and despite GS WAMI having been designated as a program of record in 2014, there is still no budget request for modernization of this combat-proven ISR system,” HASC said in its pre-conference version of the fiscal 2021 NDAA. “The committee notes that prior year congressional funding has helped GS develop beyond-line-of-sight communications, near vertical direction finding, and multi-intelligence capabilities. Additional funding is needed to modernize sensor tip and cue, sensor field of view, and to optimize machine learning in support of Project MAVEN.”
Cloud One/Platform One were to be a hosting environment for Project Maven to transform it from a developmental system into a warfighting system for the onramp.
Kicked off in 2017 with the oversight of the office of the undersecretary of defense for intelligence, Project Maven has looked to develop an AI tool to process data from full-motion video (FMV) collected by unmanned aircraft and decrease the workload of intelligence analysts who frequently spend hours sifting through FMV.
“Project Maven’s FMV Sprint 2 capability is now deployed to 40 sites within the Army, Navy, Marine Corps, and Air Force (TUAV and Medium Altitude MQ1-C and MQ9) automatically detecting and geo-locating people, vehicles and buildings, tracking objects in motion, and capturing training data for continued algorithm improvements,” the DoD fiscal 2021 budget request said. “Project Maven started by bringing AI to the FMV COIN [counterinsurgency] fight to improve the speed and accuracy of analysis and reduce the manpower burden of video exploitation. The investments in AI for FMV COIN are being leveraged for high end warfare to detect conventional military equipment such as tanks, artillery, airplanes and missile launchers. Project Maven FMV algorithms will be deployed to Air Force and Army units and additional OCONUS [outside continental United States] sites.”
General Atomics builds the U.S. Army MQ-1C Gray Eagle and the Air Force MQ-9 Reaper.
An April, 2018 Congressional Research Service report said that DoD initially incorporated Project Maven AI tools for 10 sites.
“The intelligence community has a number of publicly-advertised AI research projects in progress,” the report said. “The Central Intelligence Agency (CIA) has 137 projects in development that leverage AI in some capacity to accomplish tasks such as image recognition or labeling (similar to Project Maven’s algorithm and data analysis functions) to predict future events like terrorist attacks or civil unrest based on wide-ranging analysis of open source information.
“IARPA [Intelligence Advanced Research Projects Agency] is sponsoring several AI research projects intended to produce tangible tools for the community four to five years from completion,” the CRS report said. “Some examples of its programs include developing algorithms to accomplish multilingual speech recognition and translation in noisy environments; geo-locating images with no associated metadata; fusing 2-D images to create 3-D models; and tools to infer a building’s function based on pattern of life analysis.”
Google [GOOGL] was the prime contractor for Project Maven but dropped out in 2018 after receiving pushback from employees about the company’s tools being used for an AI drone imaging effort. California-based big data analytics company, Palantir Technologies, co-founded and chaired by venture capitalists Peter Thiel and Alex Karp, has assumed Google’s role overseeing Project Maven, according to published reports.
Early on in Project Maven, on May 20, 2017, former Deputy Defense Secretary Robert Work assigned to the Algorithmic Warfare Cross-Functional Team (AWCFT) under the Undersecretary of Defense for Intelligence the task of the automation of Processing, Exploitation, and Dissemination (PED) of tactical and mid-altitude FMV from drones in support of operations to defeat ISIS insurgents.
Big Data at War: Special Operations Forces, Project Maven, and 21st Century Warfare , an August paper for the Modern War Institute at West Point by Tufts University Prof. Richard Shultz and United States Special Operations Commander Army Gen. Richard Clarke, said that naval special warfare group teams were “first out of the gate” in using Project Maven (Defense Daily, Aug. 25).
While they “were not immediately impressed with the algorithms’ performance,” the teams “said they could see their potential,” the report said
“Later, the tools were also sent to USSOCOM’s counterterrorism teams conducting operations forward in other parts of the warzone,” according to the report. “They had analogous reactions. While the AI could place a boundary box around vehicles, buildings, and people, and display them on a map, the algorithms were rudimentary with many false detections. In this process, called geo-referencing, analysts hoped to add location information to otherwise untagged data, so they could watch and track it. But that was not possible in the startup phase. Indeed, accuracy of detections was only around 50 percent. Determining the difference between men, women, and children was challenging.”
The paper suggested that Project Maven could serve “as the springboard to prepare DoD as an institution for future wars—a transformation from a hardware-centric organization to one in which AI and ML [machine learning] software provides timely, relevant mission-oriented data to enable intelligence-driven decisions at speed and scale.”