Anyone who has played RTS games such as Blizzard’s Starcraft knows the basics of swarm warfare: you build a swarm of cheap units and hurl them against the enemy’s smaller force of more expensive units. The plan is that although the swarm will be decimated, the enemy will be exterminated. The same tactic is also used in the classic tabletop game Ogre. It pitted a lone intelligent super tank against a large force of human infantry and armor. And, of course, the real world has many examples of swarm warfare with some successful for those using the swarm tactic (ants taking out a larger foe) and some proving disastrous (massed infantry attacks on machineguns in WWI).

A modern approach to swarm tactics is to build a swarm of drones and deploy them against the enemy. While such drones will tend to be airborne units, they could also be ground or sea machines. In terms of their attacks, there are many options. The drones could be large enough to be equipped with weapons, such as small caliber guns, that would allow them to engage and return to reload for future battles. Some might be equipped with melee weapons, poisons, or biological weapons. The drones could also be suicide machines, small missiles intended to damage the enemy by destroying themselves.

While the development of military drone swarms in the United States will fall within the usual high cost of developing new weapon technology, the drones themselves can be cheap. After all, they will tend to be much smaller and simpler than crewed weapons such as aircraft, ships and ground vehicles. The main cost will most likely be in developing the software to make the drones operate effectively in a swarm; but after that it will be just a matter of mass producing the hardware.

If effective software and cost-effective hardware can be developed, one of the main advantages of the battle swarm will be its low cost. While such low-cost warfare might be problematic for defense contractors who have grown accustomed to profitable contracts, it is appealing to those concerned about costs and reducing government spending. After all, if low-cost drones could replace expensive units, defenses expenses could be significantly reduced. The savings could be used for social programs or, more likely, more tax cuts for the wealthy.

Low-cost units, if effective, can confer an attrition advantage. If, for example, it costs you $12,000 in drones to take down the enemy’s $12,000,000 fighter jet, then you stand a decent chance of winning. If hundreds of dollars of drones can take down millions of dollars of aircraft, then the situation is even better for the side with the drones. Likewise for naval vessels, land vehicles and structures.

The low cost does raise some concerns, though. Once the drone controlling software makes its way out into the world (via the inevitable hack, theft, or sale), then everyone could use swarms. This will recreate the IED and suicide bomber situation, only at an exponential increase. Instead of IEDs in the road, they will be flying around cities, looking for targets. Instead of a few suicide bombers with vests, there will be swarms of drones loaded with explosives. Since Uber comparisons are now mandatory, the swarm will be the Uber of death.

This does raise moral concerns about the development of drone software and technology; but the easy and obvious reply is that there is nothing new about this situation: every weapon ever developed eventually gets around. As such, the usual ethics of weapon development applies here, with due emphasis on the possibility of providing another cheap and effective way to destroy and kill.

One short term advantage of the first swarms is that they will be facing weapons designed primarily to engage small numbers of high value targets. For example, air defense systems now consist mostly of expensive missiles designed to destroy very expensive aircraft. Firing a standard anti-aircraft missile into a swarm will destroy some of the drones (assuming the missile detonates), but enough of the swarm will probably survive the attack for it to remain effective. It is also likely that the weapons used to defend against the drones will cost more than the drones, which ties back into the cost advantage.

This advantage of the drones would be quickly lost if effective anti-swarm weapons were developed. Not surprisingly, gamers have already worked out effective responses to swarms. In D&D and Pathfinder players generally loath swarms for the same reason that ill-prepared militaries will loath drone swarms: while individual swarm members are easy to kill, it is difficult to kill enough of them with standard weapons. In games, players respond to swarms with area of effect attacks, such as fireballs (or running away). These sorts of attacks can consume the entire swarm and either eliminate it or reduce its numbers, so it is no longer a threat. While the real world has an unfortunate lack of wizards, the same idea will work against drone swarms: cheap weapons that do moderate damage over a large area. One possible weapon is a battery of large, automatic shotguns that fill the sky with pellets or flechettes. Missiles could also be designed that act like claymore mines in the sky, spraying ball bearings in almost all directions.  And, obviously enough, swarms will be countered by swarms.

The drones would also be subject to electronic warfare. If they are being remotely controlled, this connection could be disrupted. Autonomous drones would be  less vulnerable, but they would still need to coordinate with each other to remain a swarm, and this coordination could be targeted.

The practical challenge would be to make the defenses cheap enough to make them cost effective. Then again, countries whose ruling class is happy to burn money for expensive weapon systems would not need to worry about the costs. In fact, defense contractors will presumably be lobbying for expensive swarm and anti-swarm systems.

The swarms also inherit  existing moral concerns about non-swarm drones, be they controlled by humans or deployed as autonomous killing machines. The ethical problems of swarms controlled by a human operator would be the same as the ethical problems of a single drone controlled by a human, the difference in numbers does not make a moral difference. For example, if drone assassination with a single drone is wrong (or right), then drone assassination with a swarm would also be wrong (or right).

Likewise, an autonomous swarm is not morally different from a single autonomous unit in terms of the ethics of the situation.  For example, if deploying a single autonomous killbot is wrong (or right), then deploying an autonomous killbot swarm is wrong (or right).  That said, perhaps there is a greater chance that an autonomous killbot swarm will develop a rogue hive mind and turn against us. Or perhaps not. In any case, Will Rodgers will be proven right once again: “You can’t say that civilization don’t advance, however, for in every war they kill you in a new way.”