Lethal Autonomous Weapons Exist; They Must Be Banned

> armour underneath a tank is usually much lighter, because it's difficult to hit. Missiles have to do fancy armour-penetration stuff to get through the main armour on a tank. If you can get a shaped charge stuck to the underside of a tank, you don't need to do the fancy stuff.

I'm not even sure if the air-pressures under a tank would allow a drone to fly under there. Let alone the myriad of anti-air defenses a modern tank has.

The M1A1 Abrams tank has not only the main cannon, but also a .50 Caliber M2 Browning and two M240 machine guns. In addition, there's a general expectation that a tank would be surrounded by supporting troops and equipment (and vice versa: the tank supports the troops, as per the theory of Combined Arms)

So a tank, facing a swarm of drones, probably can just fire its two machine guns and take them out. Its not like drones have any armor of any kind. Similarly, anti-personnel rifles from the supporting infantry would probably be effective against those drones.

Once we start considering weapons like air-burst grenades, I'm finding it less-and-less likely that a "10,000 Swarm" makes any sense what-so-ever. The __reason__ we mix tanks-with-infantry is because a singular weapon (ex: air-burst grenade) works only on infantry (tank armor is too thick), while a Tandem-shaped charge RPG only really works on tanks (Personnel are relatively cheap and move in groups. If you kill one, the rest of their buddies gang up and kill you).

------------

Any scenario where you're just lobbing uniformly made drones into the same area just opens them up to air-burst or machine gun fire.

-----------

EDIT: I had a lot of unnecessary words. Lets just point this out... I'm not sure if you understand how good machine guns are today.

https://en.wikipedia.org/wiki/Phalanx_CIWS

We have machine guns that can autonomously aim and shoot-down a group of missiles that travel at 1500 mph (Mach 2). And you think a swarm of 10,000 drones flying at lol 100 mph has a chance?

https://www.youtube.com/watch?v=IB8d3OaFEco

Depending on the armament: you could have up to 9000 "pellets" per shot that airburst at the expected distance to maximize the chance of hitting the target.

If a missile flying at 1500mph has no chance, why do you think a drone at 100 chance or 200 mph has a chance?

Israel's "Iron Dome" machine guns: https://www.youtube.com/watch?v=gGaqBWXM8Ko

These platforms are "portable", for a definition of portable.

https://www.youtube.com/watch?v=iAw82h-IhdQ

Free float barrels are designed more to make it so that a bipod or handgrip doesn't change the point of impact when you put wait on them, but I don't think that would matter here.

That being said, if I was designing a system like this, I'd be aiming for engagement ranges of ~100m to make target acquisition and aiming easier. You can ID something as "person" at 100m with a 4k video stream from a cell phone camera, but you'd need some really complicated optics to do the same at 1km. And at 100m, you can afford to lose quite a bit of accuracy to recoil.

I agree 100%. So naive. Just think of a world where we (US, EU, etc.) stop investing in this technology while Russia, China, etc. continue.

It really burns me to hear AI professors rail against “autonomous weapons” yet go work at Facebook, Google, etc. Information, especially manipulation of it, is a weapon too.

I think there's a big difference between a large, hot, predictable missile and a tiny, cold, evasive drone.

Area-effect weapons would have more impact, sure. But that's a lot of area to cover. A bit like flak vs airplanes - you've got to shoot a lot of metal into the air hoping for a bit of shrapnel to hit something vital.

It's the "big and expensive" vs "tiny and very cheap" decision - you don't mind losing 90% of your swarm to take out one tank, because the entire swarm costs less than the tank (added bonus: there's no humans in the swarm).

I agree with much of what you say, but I think you're limiting your imagination in drones' current and imminent capabilities are.

Drones can, and are, replacing fighters and cruise missiles, if not make those ideas of weapons platforms entirely obsolete. Autonomous jets exist already, you may have not heard of missiles composed of thousands of small single-use bombs with onboard auto guidance, which also exist already.

While yes, loitering is a strength, it is far and away not the only thing a robot can be engineered to do.

The bigger idea here is the problems being solved when diplomacy breaks down into violence. You seem to not be aware that modern warfare consists of Shock and Awe. Not like the movies. If there could be a combatant in an office building in the next village, you call in an AC130 and you blow up the building. The idea is that warfare isn't a battle of contrition, but a battle of financial resolution.

So, looking at it from this angle, yes; the natural evolution of warfare will be robot on robot until one side runs out of the ability to make more robots.

Everything else resulting from this has been stated already.

But I think you're just underestimating human engineering and imagination.