This post is also available in: עברית (Hebrew)
The US Defense Department is getting assistance from Google’s open-source AI software. The search giant’s AI will be helping the agency to analyze defense footage. Google’s pilot project with the Defense Department’s Project Maven, an effort to identify objects in drone footage, has not been previously reported, but it was discussed widely within the company last week when information about the project was shared on an internal mailing list, according to sources who asked not to be named because they were not authorized to speak publicly about the project.
Project Maven, a fast-moving Pentagon project also known as the Algorithmic Warfare Cross-Functional Team (AWCFT), was established in April 2017 in order to “accelerate DoD’s integration of big data and machine learning.” The project’s first assignment was to help the Pentagon efficiently process the deluge of video footage collected daily by its aerial drones. Although the Defense Department has poured resources into the development of advanced sensor technology to gather information during drone flights, it has lagged in creating analysis tools to comb through the data.
Maven was tasked with using machine learning to identify vehicles and other objects in drone footage, taking that burden off analysts. Maven’s initial goal was to provide the military with advanced computer vision, enabling the automated detection and identification of objects in as many as 38 categories captured by a drone’s full-motion camera, according to the Pentagon. Maven provides the department with the ability to track individuals as they come and go from different locations.
A Google spokesperson told gizmodo.com in a statement that it is providing the Defense Department with TensorFlow APIs, which are used in machine learning applications, to help military analysts detect objects in images. “We have long worked with government agencies to provide technology solutions. This specific project is a pilot with the Department of Defense, to provide open source TensorFlow APIs that can assist in object recognition on unclassified data,” the spokesperson said. “The technology flags images for human review and is for non-offensive uses only. Military use of machine learning naturally raises valid concerns. We’re actively discussing this important topic internally and with others as we continue to develop policies and safeguards around the development and use of our machine learning technologies.”
Some Google employees were outraged that the company would offer resources to the military for surveillance technology involved in drone operations, sources said, while others argued that the project raised important ethical questions about the development and use of machine learning.
In any event, it’s possible that Google’s own product offerings limit its access to sensitive government data. While its cloud competitors, Amazon and Microsoft Azure, offer government-oriented cloud products designed to hold information classified as secret, Google does not currently have a similar product offering.