Article published in Defense News
Last month’s use of a Predator drone equipped with Hellfire missiles to kill Al Qaeda operative Abu Ali is being hailed by transformation proponents as further proof that future warfare will favor unmanned combat vehicles. That view has become increasingly popular among policymakers, who see unmanned vehicles, especially aircraft, as a safer and less expensive way of conducting combat missions traditionally assigned to manned systems.
Before making a big bet on unmanned systems, though, the policymakers might want to consider another news story that appeared in November. That story reported that the Department of Energy is planning to build the world’s most powerful supercomputer, a behemoth capable of crunching numbers at the rate of 100 trillion operations per second.
What’s the connection with unmanned systems? A hundred trillion operations per second is the rate at which the human brain operates. In other words, the most powerful computer in history will be required to match the computational capacity of a normal person — like a pilot.
Don’t expect to see that kind of machine intelligence on an unmanned vehicle anytime soon. The new computer, code-named “ASCI Purple,” will take up space equivalent to two basketball courts and require 119 miles of optical wiring. An unmanned vehicle hosting such hardware would probably look something like the Hindenberg, and be about as survivable.
Of course, the current generation of unmanned aerial vehicles depends on links to remote human pilots to be viable, rather than presuming to match human performance with onboard logic. That’s one of the reasons why a third of the Predators have been lost — when tight maneuvers cut the drone’s data links or transmission delays slow the receipt of time-urgent situational information by remote operators, Predators have a tendency to crash.
The heavy dependence of current and prospective unmanned vehicles on remote human operators should raise a red flag of warning for those who think the Pentagon is about to enter the age of robotic warfare. Unmanned systems such as Predator and Global Hawk offer unique capabilities to warfighters such as day-long endurance over contested areas, but they are too fragile to conduct most combat operations.
Such limitations tend to get lost in the public debate of military transformation. The reluctance of military leaders to embrace unmanned vehicles is often attributed to lack of vision, or unwillingness to challenge bureaucratic constituencies. Pilots don’t want to give up their airplanes and flight pay, the critics contend.
That complaint infuriates service representatives, most of whom view unmanned systems as offering niche capabilities rather than revolutionary warfighting potential. In a recent high-level meeting where he was getting the hard sell from policymakers on a particular unmanned system, one senior officer finally exclaimed, “If this vehicle had a person in the cockpit, we wouldn’t be talking about it because everyone here would realize it can’t do much!”
For many skeptics, that is the crux of the problem. As a senior political appointee puts it, once you get past the “dog-dancing-on-its-hind-legs novelty of unmanned combat vehicles, they aren’t very capable.” That complaint doesn’t apply to long-endurance surveillance drones, which have a well-defined role that really does add to the capabilities of existing air-breathing and orbital assets. For example, the Navy thinks it can save money and manpower by assigning routine maritime surveillance missions to the high-flying Global Hawk vehicle.
But when it comes to conducting actual combat operations like close air support or defense suppression, the limitations of unmanned vehicles are all too obvious. Consider November’s Predator strike. The air space was uncontested. Nearby ground spotters directed the Predator to the target. The Qaeda operatives had no defensive sensors or missiles, and no capacity to jam data links.
If these very benign conditions had not prevailed, the strike could not have occurred. Even the most dated and decrepit tactical aircraft can shoot down a Predator. Without ground spotters, targeting the terrorists’ vehicle would have been very difficult. If the terrorists had active or passive defenses, the attack could have been foiled easily.
That doesn’t mean that unmanned vehicles can’t be made more robust. For example, the Air Force has launched a program to equip Predator with Stinger missiles so it won’t be as vulnerable to attack. Future vehicles can be made more stealthy, and redundant data links can be installed. But that will take time and money, and even after it is accomplished the vehicle will not begin to approach the resilience or versatility of a manned aircraft.
What the military services need is not novelty but new capabilities — the kind of transformational capabilities that enable new concepts of operation. Unmanned systems may be able to provide that boost someday, but it would be premature to begin eliminating manned systems based on what is known today. The technology is simply too primitive.
In 1921, playwrite Karel Capek coined the word “robot” to describe a mechanical man that seemed to be within the grasp of contemporary technology. Eighty years later, you still don’t run into many robots on the subway. Artificial intelligence became a technological dead-end, because the human brain turned out to be more complex than anyone imagined.
That’s still true today. No unmanned system can match the analytical skills of a human warrior. Even the Energy Department’s new supercomputer will have only half the hundred trillion bytes of memory in a typical human brain. So whatever warfighting vehicles we build will have to be closely tethered to a real person, and heaven forbid they should encounter a determined adversary who actually is human.
Loren Thompson is Chief Operating Officer of the Lexington Institute.
Find Archived Articles: