11/8/2015
Pentagon Will Develop ‘Thinking Machines’ to Defeat Future Enemies
By Sandra I. Erwin
SIMI VALLEY, Calif. — Ultra-smart computers and robots that crunch data at the speed of light will be key ingredients of the Pentagon’s strategy to deter and defeat future military adversaries.
“We already started to make investments,” said Deputy Defense Secretary Robert Work.
The specifics of how the Pentagon might use these advanced technologies and how they will be acquired will not be known for years, Work said, but the Pentagon today is convinced that using computers to aid decision making is a “big idea.”
Autonomous computers that can collaborate with human brains will be central to what the Defense Department calls its “offset strategy” to ensure technological dominance for the coming decades, Work said Nov. 7 during a talk at the Reagan National Defense Forum at the Ronald Reagan Presidential Library. He threw around catchphrases like “human machine collaboration” and “combat teaming” to describe this vision of the future.
Work for years has called on the Pentagon to make ambitious and bold investments to create technological “surprises” as the United States did during the Cold War, when defense planners figured out how to “offset” the Warsaw Pact’s much larger conventional forces with nuclear weapons. That advantage did not last, though, as the Soviet Union quickly moved to build its own nukes. The next wave of innovation came in the 1970s when Secretary of Defense Harold Brown and Undersecretary William Perry pushed a new offset strategy built around the use of digital microelectronics and information technology to counter conventional forces. The result was a wave of innovation in smart weapons, sensors and command-and-control networks. Work credits the second offset for propelling the United States into unchallenged superpower status.
The third offset will be about machines that can learn and operate “literally at the speed of light,” he said. The tech industry already has created supercomputers that can beat human chess champions, and the Pentagon plans to ride that innovation wave into the next level of autonomy and artificial intelligence technology.
The idea is that only “learning machines” can keep up with quickly morphing threats like cyber or electronic attacks, and react far more quickly than human brains ever could. Computers also would help military commanders figure out how to respond to incursions against U.S. assets in outer space or massive hypersonic missile strikes. “You need machines to help you solve problems right away,” Work said. “The way we’ll use machines is to help humans make better decisions faster.”
This is a significant departure from the way defense and deterrence strategies were developed in the past, he explained. “This assumes that an adversary changes strategy all the time.”
An illustration of how the Pentagon increasingly values advanced data analysis — as a capability that must go hand in hand with conventional mega weapons — is the F-35 joint strike fighter. “The F-35 a flying sensor-computer that sucks in an enormous amount of data, correlates it, analyzes it and displays it for the pilot on his helmet,” Work said. Legacy fighter jets like the F-16 might be as good or better than the F-35 at traditional aerial dog-fighting, but “we are absolutely confident that the F-35 will be a war winner, because it helps make better decisions.”
Another tenet of the third offset strategy is to insert robots more aggressively into combat operations, Work said. “If 10 years from now, if the first person through a breach isn’t a frigging robot, then shame on us,” he declared. “Secretary [Ashton] Carter is absolutely certain we can do this.”
Work cautioned that the third offset is not purely a technology game, but also will require unconventional concepts for how forces operate and organize. Having a “vibrant” mix of military, civilian and contractor innovators also will be essential, he added. Carter has been insistent that the Pentagon recruit startups and commercial tech firms in Silicon Valley that are on the cutting edge of big data and robotics applications.
“It’s going to take some time” to bring this to fruition, Work said. “There will be a period of experimentation.”
Human-machine collaboration will not only be used to fight major wars but also low-intensity conflicts. Today’s counterterrorism campaign, for instance, is a network of computers, drones and special operations teams that work with massive amounts of data. In the future, Work said, “You don’t go after an automated solution. You go after big data analytics.”
In the face of Islamic State militants who flood the Internet with millions of social-media posts, for instance, advanced machines would be able to crunch the data, and suggest “how we might be able to go after them.”
Defense officials are laying the groundwork of the third offset strategy, with the expectation that future administrations will continue the effort, which could take decades. “We want to have the intellectual underpinning to set us on this path,” said Work. During the remainder of the current administration, the plan is to conduct demonstrations and experiments. “The transition from administration to administration over a long period of time is based on a commitment that this works, that it gives us an advantage.” For the foreseeable future, “We’ll do more demonstrations than procurements. We won’t be able to buy capabilities as fast as we’d like. We want to make sure we keep pace with potential competitors’ capabilities.”
http://www.nationaldefensemagazine.org/blog/Lists/Posts/Post.aspx?ID=2011
