The Pentagon’s Strategic Capabilities Office is focusing more on software than hardware, and trying to pivot the department to becoming a data-driven organization, SCO Director Will Roper said Monday at the SXSW Conference in Austin, Texas.
In a wide-ranging conversation about his office’s work and the future of warfare, Roper answered a question around how to handle the department’s newest weapons potentially falling in the hands of the Islamic State.
“We need to do a better job in the Defense Department of not over-designing things so that we can’t lose them,” Roper said.
He said the DOD is good at designing exquisite systems, which use the best of government technology, but when it comes to autonomous weapons the department is trying to stay within the realm of commercial technology. While it may not be as high-performing as something highly-customized and built in-house, if its left behind on the battlefield there is not as much worry it could be turned on the U.S.
Instead of focusing on hardware, then, SCO focuses on investing in the software underlying those commercial technologies that allows systems to collaborate, Roper said. And software, he said, is easier to protect than hardware.
His comments come as a recent report from the Pentagon’s science advisers said most of the U.S. military’s weapons systems were built without any effort to protect them from cyberattacks on hardware components. There is evidence that some already have been fitted with digital backdoors, meaning an enemy could make them fail in a real conflict, the report from the Defense Science Board said.
‘Shift it to being data-centric.’
Also asked what keeps him up at night, Roper responded: “I think the Pentagon’s lack of recognition that data is going to be one of the pioneering tools” for the future of warfare.
He continued: “We don’t treat data the same way companies like Google, or Apple or Amazon do. For us, data is kind of like an exhaust that comes out of our systems.”
The future of warfare, Roper said, is to use data like a strategic resource. That means thinking the way private companies do about artificial intelligence and machine learning, and how data can be used to train machines.
“This is a culture shift, so this is tough to do, is try to take a Pentagon that is device-centric,” he said, “and shift it to being data-centric.”