The U.S. government wants to use more “Big Data” analytics and crowdsourcing to help keep tabs on the proliferation of nuclear weapons, according to a new report from the Department of Defense’s Defense Science Board.
That report (PDF), titled “Assessment of Nuclear Monitoring and Verification Technologies,” details the capability of the Department of Defense, Department of Energy, and the U.S. intelligence community to monitor nuclear activities around the world. Those agencies aren’t just concerned about existing nuclear powers increasing their weapons capabilities; they’re also kept awake by the idea of states and non-state actors possessing (or attempting to possess) nukes of their own.
“The technical approach for monitoring cannot continue to derive only from treaty and agreement dictates for ‘point’ compliance to the numbers and types formally agreed upon and geographically bounded,” the report explains in its introduction. “Proliferation in this future context is a continuous process for which persistent surveillance, tailored to the environment of concern, is needed.”
The researchers behind the report recommend the latest generation of analytics tools as a way to chew through enormous amounts of structured and unstructured data from sensors and other sources, the better to determine whether a country is building a nuclear weapon or improving its existing capability.
In a somewhat unusual twist, the report also advocates crowdsourcing, particularly of commercial imagery, as a way of ferreting out nuclear weapons under development. It cites the example of a group of Georgetown University students who relied on Google Earth, online photos posted by Chinese citizens, and other imagery to analyze a network of tunnels in China used to hide part of that country’s nuclear arsenal; they used that data to publish a controversial paper that challenges the U.S. government’s assessment of Chinese nuclear weapons capability.
“Whether the [students’] report is completely accurate or not,” the report suggests, “this event provides a ‘proof-of-concept’ on how crowd-sourcing can be used to augment limited analytical capability.” By establishing a process that “codifies” crowdsourcing as an area of supplemental research to nuclear-treaty-monitoring issues, it adds, “open source and mainstream intelligence assessments can be reconciled.”
The use of increasingly powerful analytics (and crowd-sourcing) seems like a natural direction for something as complicated as nuclear-weapons monitoring to take. But even with the most advanced tools at the government’s disposal, ambiguities will surely remain, if only because human beings (with all their faults, follies, and advantages) will ultimately need to serve as the ultimate arbiters of what’s germane to a particular investigation.