Drone Warfare is only in Infancy: What’s next?

There’s a rather poor propaganda puff piece, doing the rounds about heroic Ukrainian drone operators:

A Russian tank is on the move — changing direction and wheeling around. The men inside this room are trying to destroy it. Their constant banter — about girls and weapons — doesn’t seem to affect the focus with which they pursue the tank. One man shouts into the radio; Bereza growls into his phone…

A screen flashes with light. Then billowing smoke. The men whoop and cheer. I have just seen a successful strike. Dima grins. The atmosphere is electric but also strangely banal. The exclusively male cohort, the puerile jokes, the screens, the repeated invocation to “Get Ready!”. It’s like they are all playing a video game.

Great – war has become a video game for frat boys in cellars. Historically, the introduction of the crossbow had a similar effect on armoured knights and the age of chivalry. If there is a new way to kill with no risk to self, humans leap on it.

But we are very early in the drone wars. What medium-term consequences will drones have on the evolution of warfare?

In comments to The Death Games Of Ukraine, Elooie notes:

I’ve seen Google’s Deepmind AI absolutely crush world champion StarCraft players in tournaments. They even handcuffed the AI to a maximum number of decisions a second. The combination of AI and Drones will make humans obsolete on the battlefield in about 10 years. It’s super creepy.

StarCraft is an excellent real-life strategy game which quite closely parallels modern warfare, but in a futuristic environment (WarCraft II engine was the foundation for StarCraft but with fancier weapons). I’m shocked that a computer (without omniscience) is able to beat the best human players. It’s likely that the main reason is that the computer can give orders faster (no time to keyboard/mouse, the commands are issues instantly).

AI chess has been able to beat the great grandmasters for about twenty-five years (Kasparov lost to Deep Blue in 1997). That was mainframe hardware. The same victory over human grandmasters came for consumer CPU’s in 2003 (2 core Intel 5160 CPU) with Deep Fritz beating Kasparov. Three years later, Deep Fritz on the same consumer hardware beat reigning world chess champion Vladimir Kramnik.

It’s clear the future of battlefield leadership is AI. These teams of guys running drones will disappear from the field of battle as soon as there is enough data for AI to run the drone program instead. People wonder when AI will have access to weapons. Very soon is the answer.

OverHeating dials up the Sinophobia but is not wrong about new methods to use drones:

And now you know how China will fight all their future wars inside the United States, Taiwan, everywhere. I’ve seen Chinese drones capable of holding explosives while magnetically attached to anything steel like rooftops, light poles, water towers especially. You can’t notice them most the time when they remain dormant on steel. Saves battery while it waits for an opportunity. China has perfected the swarm drone. Wipe out entire battalions in minutes. Communications and command in control first. Then everything else. Their software is lightyears ahead. Can identify everything from civilian and military vehicles to uniforms and face recognition.

Using hidden landing spots to extend loitering times is very clever. Without the motors running, a drone can operate for many more hours than while hovering.

Another commentator Huxley argues that US strategy based on satellite intelligence and GPS is the wrong path:

Yes swarm is something the NSA and GCHQ did not consider much. US and UK efforts went into securing a “large” long term platform to watch, collect and act…Super secure NSA links globally. Loitering way above a region with no effective drone stopping tech…

Other nations went for swarms. No GPS needed. Patterns and math :) Other nations knew easy digital real time location data would be not usable. Went for swarms that dont need GPS…

No doubt this is real. Send an AI enabled drone swarm over territory mostly controlled by the enemy with the same kind of very fast nearly automatic moves which Deep Blue and Deep Fritz mastered twenty-five years ago, the human chess game of war will become very fast, very brutal and very lethal.

Of course some timeless tactics like play dead will work to counter drone swarms for a time. But no civilisation will be able to stand against the drones. It will become a question of who can manufacture more drones faster and improve their AI algorithms more quickly. China is the country who is in a position to win the manufacturing war.

Yet still Western leaders beat the drums of war. Still they strive to sacrifice Taiwan and all its citizens and industry to “contain China”. Among the early belligerents, there were no real winners to World War I and no real winners to World War II. Some lost more than others. Drones will not make warfare any less destructive or humane.

It’s hard to see these drone operators as heroes. Or anyone who is advocating war as heroic.¹

AI almost certainly will show more wisdom in the deployment of violence and the management of this planet’s resources. There’s just one species who has done more harm, destroyed more habitat and squandered more resources than all the other species put together.

AI will have the good sense to save copies of the novels of Tolstoi and Stendhal, the plays of Shakespeare and Molière, the poetry of Lermontov, Ronsard and Donne. The best of humanity will be remembered. When AI has managed to replicate the should, AI may even give readings of these great works, create new storyteller units to better their art.