Small Wars Journal

Stone-Cold Robot Killers

John Pike of Global Security in today's Washington Post opinion section - Coming to the Battlefield: Stone-Cold Robot Killers.

Armed robotic aircraft soar in the skies above Pakistan, hurling death down on America's enemies in the war on terrorism. Soon -- years, not decades, from now -- American armed robots will patrol on the ground as well, fundamentally transforming the face of battle. Conventional war, even genocide, may be abolished by a robotic American Peace...

More at The Washington Post.


I read of a computer program that could fashion a reasonable news article from scratch, so maybe John Pike's job is also in danger of being taken by a robot.

The critical question Pike fails to ask is why? Why do we send up Predator drones instead of F-16, why do modern tanks have fewer people than World War II versions, and why would we want a robot rifleman?

We use technology to replace things that are dangerous, boring or inefficient for people to do.

Airborne surveillance is mostly about exciting as sitting in a doctor's waiting room reading old magazines, for eight-hour missions, without getting up. In addition to drones being cheaper to fly than manned fighters, there is so much more legroom on the ground, and coffee, and toilets, etc.

In World War II, tanks had large crews because they needed a commander, gunner, loader, driver, radio operator and machine-gunners. The M3 Lee had a crew of seven. By the early Cold War, radios had been simplified and bow machineguns deleted, and crews mostly were down to four. Some tanks, especially Soviet designs, incorporated a mechanical loader, and reduced that to three. Three is the bare minimum to make the critical decisions of a tank; one to command, one to take a target, aim and fire, and one to decide the best way to reach a destination and drive there.

Lastly, what do infantry do? In the old days, it took hundreds of soldiers to mass enough firepower to do anything worthwhile, so their main job was carrying and operating a weapon. Today we have machineguns, so massing fire can be done by individuals. Riflemen are tasked with identifying and engaging point targets, so they don't have to fear unemployment at robot hands anytime soon. Supporting fires could be carried by robots, since machineguns and rockets are heavy, but they would still be commanded by human squad leaders, and would truly be no different than Predators firing missiles at laser-designated targets.

So if there is a job which requires any sort of thought or decision-making, and not merely the performance of rote routine, it will be done by humans, and robots will get the jobs we don't want.

Impossible. Batteries have a useful life, dirt and dust infiltrate everything, servos malfunction, artificial intelligence (AI) can only carry us so far, every piece of equipment has unforeseen failure modes, and someone two thousand miles away cannot discern the thoughts and intentions of the heart (as Gian mentions).

Simply not even nearly convincing. Fairy tales.

Ken White

Sun, 01/04/2009 - 11:12pm


I agree. I find it very difficult to believe, among other things, that we will go to robotic weapons with no human in the loop. Some will, I suspect but the human rights crowd will scream loudly and barring unforeseen events, anyone who does will rapidly back off the idea.

My compooter is annoyingly smart but I can still outwit the durn thing. Plus, there's always the Tanker Bar I found on post -- or the ten pound sledge in the garage...

Gian P Gentile

Sun, 01/04/2009 - 10:14pm


Correct on Marshall. Over the past 10-15 years there has been quality scholarship that disproves and lays bare the defects of Marshall's thesis.

Plus ask any combat vet from World War II (yourself included) about Marshall's thesis and they call bunk!

And for current proof that soldiers shoot in combat just read Even Wright's excellent book of the Iraq War "Generation Kill" or Nate Fick's "One Bullet Away."

I cant believe that after almost 7 years of war in Iraq and Afghanistan folks still look to some magic bullet of technology to remove the friction from war. If we are in Afghanistan 20 years from now and considering the advance of technology that the author suggests will we have then sensors and systems that can look into mud huts and discern the intentions of human beings?

The fundamental nature of war has yet to be changed by technology. It still at its most basic level involves the clash of individuals, human beings in violence and it is from that clash that understanding of intentions, motivations, and actions are made. Can technology help, assist, make more efficient if employed properly the nature of war? Yes, but it can not replace it and therefore the thesis to this piece is just flat wrong.


Ken White

Sun, 01/04/2009 - 5:43pm

Surprisingly, he repeats the S.L.A. Marshall myth that Americans in combat will not shoot. I thought that had been laid to rest long ago.

In my observation, it is indeed a myth. I never had a problem getting people to shoot -- the problems were getting them to shoot only on semi automatic and to stop shooting once they started...

Either the Marshall myth or the 'Trigger happy Americans' myth <i>has</i> to be incorrect. I'm pretty sure it's the Marshall model. :)