Kansas City I can still remember my third grade teacher walking us through the duck-and-cover exercise and announcing as we prepared to march to the corner single file that she would be standing on the asphalt covered playground where she intended to mark an X, because she wanted no part of surviving a nuclear holocaust. Being raised in the Sputnik era in the heat of the Cold War, what could possibly scare me? The answer is a blood-curdling, hair-raising article in Scientific American about our slim chances of surviving or at this point even stopping the pell-mell rush of a number of countries towards automatic, algorithm-driven warfare.
The piece was written Noel Sharkey, a professor emeritus at the University of Sheffield in England. Sharkey is also the founder and chair of the International Committee for Robot Arms Control and one of the co-founders of a nonprofit called the Campaign to Stop Killer Robots. In short, he’s not neutral on this issue, thank goodness!
Automatic weapons systems (AWSs) are not simply drones, which are scary enough even though nominally under the control of joystick jockeys here and abroad. He actually advocates maintaining a human touch on any and all weapons both at the trigger and, more importantly, as a last gasp flesh-and-blood safety against unstoppable disaster. He’s worried not about Star Wars cloned robot weapons, but about computer controlled tanks, planes, submarines, and ships that have a mind of their own in the deep binary of computer code and algorithms programmed to find their targets, engage equally mindless “enemies,” who may be other unmanned tanks, planes, submarines, and ships caught in the same endless kill and destroy loop.
It sounds like science fiction, until you realize that the United States is already developing unmanned transoceanic ships with fighting power that are accompanied by what Sharkey refers to as a flotilla of unmanned submarines that also have fire power. They even have names, Sea Hunter for the ship, and DASH (Distributed Agile Submarine Hunting) for the subs. Of course, the USA is not alone. Russia doing the same with unmanned tanks designed to station along their border to stop invasions, like a weaponized Trump border all. China, Iran, and other countries are all involved in developing similar weapons.
AWSs is the mostly deadly horror story of what might be wrought by artificial intelligence and machine learning, even as all of the experts concede computer glitches happen and software fails, even as the claim that they are preparing fixes for that. I’m not sure what might stop hacking, since there seems to be no stopping that in normal life, much less on a battlefield.
Sharkey tells a story of how badly this can go by citing an algorithmic faceoff between two bookseller websites, profnath on one side and bordeebook on the other in a fight in 2011 on the Amazon website over an out-of-print book. It was usually offered for $50, but “every time bordeebook increased its price, so did profnath.” “Within a week brodeebook was selling the book for $23,698,655.93 plus $3.99 shipping before anyone notice.” His point was hard to miss. Similarly programmed automatic warfare instruments designed to act and react to expected battlefield or war scenarios could also careen into a totally destructive abyss. The difference is that it wouldn’t take a week before the impacts were known, though it might take longer to count the bodies and forever to repair the damage.
Thirty nations have called for a complete ban of fully autonomous weapons according to Sharkey. Most other countries want at least a guarantee of human intervention in the decisions to attack or not. “Progress is being blocked, however, by a small handful of nations led by the U.S. Russia, Israel, and Australia.”
Tell me this doesn’t frighten you. Literally to death!