Lethal Autonomous Weapons Exist; They Must Be Banned
spectrum.ieee.org
spectrum.ieee.org
Britain and France, notably, hadn't really worked out how to use them effectively and so were left fighting the last war, horribly disadvantaged until they worked out how to catch up (though air power being developed at the same time also had an influence).
So now we have drones, and all those incredibly expensive fighter planes are useless. Aircraft carriers will be replaced with cheap drone swarm carriers, so naval strategy will need to change. Anti-tank drone swarms will make conventional armour more or less ineffective. And so on.
The good news is that there will be anti-drone swarms too. And so we'll end up fighting battles mostly by destroying drones instead of people. Once you've knocked out all the enemy drones, they'll surrender because humans can't fight drones.
Interesting times.
Humans aren't that logical.
Qassam missiles are straight, dumb ballistic missiles with no guidance.
I think they're perfectly logical. Superior force (in the context being discussed) makes you better at war, but not ruling.
Advances in military tech will probably reduce the overall bodycount of wars, but they don't solve the insurgency problem.
1. This won’t be your average <$1000 quad copter with a two kilometer range
2. Even if it was it could be hundreds of them
3. The military are developing large aircraft that function as essentially fighter aircraft that can destroy your tiny gun battery from 50-100 km away
Seems to me the likely scenario is some sort of cheap projectile automatic weapon mounted on a vehicle. It doesn't take much to take down a drone and they move incredibly slowly compared to "real" missiles. Sure, 22LR, why not.
I don't think any kind of conventional gun is going to do anything against this.
Also: Soldiers already wear armor designed to stop bullets. Not sure this is an obvious gamechanger.
Because they're subsonic soon after they leave the barrel. A swarm could have fairly precise directional hearing, so dodging such a slow bullet isn't out of the question.
But the real issue is that the threat is a swarm, why are you taking potshots at it. Whatever weapon you use should be able to consume as many of them as possible at once.
Then the enemy will build, or buy, more drones. Well-funded armies or terrorist groups don't lose because they run out of weapons.
Where bots and swarm bots would be used is likely in asymmetric warfare.
"Aircraft carriers will be replaced with cheap drone swarm carriers" I would argue that Aircraft carriers are the size and cost due to large munitions and maintenance constraints.
also, drones can do stupid things like fly underneath tanks, or land on them and then place a charge right up against the armour (or a vision slit, etc). Or even place a charge, back off, detonate it, then place a second charge on the same spot and detonate that. And of course a swarm can see multiple drones attack exactly the same spot on a tank's armour.
also, they don't need to destroy the tank. They just need to mess with it enough that it needs repair.
I'm not sure how this is superior to just shooting a tank with a missile.
> Or even place a charge, back off, detonate it, then place a second charge on the same spot and detonate that
Now you've halved the payload. If your drone can carry 10kg, you now have 2x 5kg bombs instead of 1x 10kg bomb. If you have a single payload, you have just a shitty, slow moving cruise missile.
> And of course a swarm can see multiple drones attack exactly the same spot on a tank's armour.
We have *guns* that attack the same spot on a Tank's armor. Let alone missiles or drones. Al Qaeda was chaining RPGs in the last war together (__literally__ a chain) that accomplishes that goal.
--------------
The __ONLY__ advantage I can see drones having is sustaining a position for hours at a time. Missiles always fly at top speed, so you can't just "wait around" an area with a missile.
A drone with a gun can sit in a location for 4 hours waiting for the tank to come, then fire the RPG when the tank comes into position. That is, the drone plays a role similar to a modern infantry, except the drone is cheaper to make.
But as soon as you're talking "hit the same spot twice", then we're back to guns / RPGs with a chain on it. Maybe a drone can fire that gun, but... don't have the drone do the job directly.
---------
EDIT: The theory of "Combined Arms" means that in practice, you'll never face a lone tank. In practice, infantry move to protect tanks, and tanks move to protect infantry.
If you are a soldier who is facing an enemy army: you only have enough strength to carry one weapon. So what do you pick? An anti-tank RPG? Or an anti-infantry machine gun?
Drones seem to fit the same conundrum on both sides. I'd expect drones to be weak against small arms (a machine gun would quickly disable a drone, even if it moves at 100 mph). While RPGs are useless vs drones. Similarly, a drone has a very small payload, so it only can have a certain number of weapons.
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.
> If you have a single payload, you have just a shitty, slow moving cruise missile.
yes. exactly. Except that for the cost of a single conventional cruise missile, you now have 10,000 shitty slow-moving, intelligent, able-to-do-evasive-maneuvres missiles.
> Al Qaeda was chaining RPGs in the last war together
so it works then? Imagine 10,000 RPG grenades flying by themselves, able to chain-hit a target, with no human operator in sight.
> A drone with a gun can sit in a location for 4 hours waiting for the tank to come, then fire the RPG when the tank comes into position.
The drone is the grenade. It doesn't fire the grenade. It flies to the target, attaches itself, goes bang. Next one flies to the same point, attaches, goes bang. There's thousands of them.
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.
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).
Apparently this hypothetical drone you're talking about has the evasion of a missile, and the stealth capabilities of the latest $Billion stealth aircraft, while 10,000 of them put together costs less than a single M1 Abrams tank.
It can also fly at speeds to evade our aimbot CIWS or aimbot CRAM defensive guns and carries enough of an explosive payload to damage an M1 Abrams.
Color me skeptical. I find it unlikely that you can make such a device for $1000ish.
By their nature, a drone could be harder to detect than another threat. A small drone will have a small radar return, and battery powered electric motors will generate less of a thermal signature than a rocket motor.
I'm not at all saying that a CIWS or CRAM would be totally ineffective against a drone or a swarm of drones, but that they may not be the impenetrable shield, either. There are limitations to sensors, target classification, rules of engagement (will the CRAM automatically light up a drone with 50 rounds if the backstop is an office building or residential block?), and ammunition capacity.
Let's take CRAM, AFAIK they are mostly installed to protect against indirect fires, so that may imply that they're tuned/configured against certain incoming projectile trajectory, speed profiles, and radar returns. Bird like radar signatures may be filtered out, or level flight trajectories that do not indicate an impact within the protected area. Nothing that can't be changed, of course, but such change might come after a few bloody engagements.
In a true future war scenario it wouldn't be one hobbyist drone with a jury rigged mortar round against a CRAM installation either, I would expect there to be drones flying around blasting out wide-band EM hash, dropping lines of chaff, and using other measures to clutter the battlespace and confuse and overload defending sensors to allow offensive devices to get through and damage or destroy their targets.
Unmanned Predator drones shoot Hellfire missiles at targets.
Stealth RQ-170 Drones stalk the enemy and obtain intelligence without tipping off the enemy.
Maybe other drones exist (maybe so stealthy that their existence is still unknown to the public), but we already know of these two strategies from public reports in Afghanistan / Iraq.
--------
We also have models for what we expect the next 10 years or 20 years of (potential) war to look like. A likely scenario is that China / Taiwan relations sour dramatically, and China decides to attack Taiwan for some reason, and the US rushes to Taiwan's aid.
Under such a scenario, the USA would be armed with whatever our Carrier Strike Groups can carry (CIWS, Cruise Missiles, a myriad of supporting destroyers and submarines, multiple air-wings, etc. etc.). While China would be armed with its huge Air Force, its increasingly huge Navy (of mostly small ships, but a few smaller carriers are on the way), and most worryingly: China's high production output can build huge numbers of cruise missiles.
----------
There's probably a few other locations in the world where USA vs Russia could get into a ground conflict. But... lets take the China vs Taiwan scenario for a sec and think.
Where does drone warfare fit? Stealth Drones for detecting enemy movements would be key: both for US and China. Carriers are fast and expensive: China absolutely would want to neutralize them. But how does China find a group of sea-vessels traveling at 40-knots in the first place?
And that's assuming that those drones aren't shot out of the air by anti-air defenses.
---------
Maybe the detection problem is solved, and now China is ready to fire upon a US Strike Group (not necessarily at the carrier, but maybe at its anti-air destroyers / cruisers in preparation for future strikes).
It seems like the obvious weapon of choice would be China's hypersonic cruise missiles: traveling at Mach 5 and with coordinated swarming AI, its the obvious way to avoid CIWS / CRAM / Patriot missile defenses.
Once the anti-air defenses are down, China can continue its attack with cheaper subsonic missiles to defeat the rest of the fleet. No drones required, aside from maybe the initial recon missions.
-------------
As far as people can tell: it seems like it will be a game of cat-and-mouse. If the US Carrier Strike group can avoid detection, the US would win the combat.
But if the US Carrier Strike group is detected: then its all over. China has more than enough cruise missiles to win.
--------
I'm not seeing how, or where, a swarm of cheap electric-motor autonomous quadcopters comes into this playbook. That's bad for stealth / recon, because the enemy knows you're watching them. (AKA: The Aircraft Carrier will just launch an F-22 and shoot down the swarm, blinding the would-be intelligence). To be successful at its recon mission, the drones have to secretly watch and pinpoint the carrier strike group, or survive a dedicated attack from the Aircraft Carrier's myriad of squadrons at its disposal.
If we're already past the detection phase for some reason, then there's no need to use drones. I think the expectation is that China's Mach 5+ hypersonic cruise missiles will instant-win the combat. Those move too fast to be shot down reliably. (Patriot missiles only fly at Mach 4)
Or maybe, the current state of the art of missiles, is superior to that strategy?
Consider 1 hypersonic missile, 1 supersonic missile, and maybe 10 subsonic missiles, all cruise missiles flying no higher than 100m off the ground.
Since they're so low, they can only be physically detected at a range of 10km (they're literally "behind the horizon" and cannot be detected beforehand).
With proper timing: all 12 targets pop up on the horizon simultaneously. The 10 subsonic missiles will hit the ship in 30-seconds, the supersonic missile will hit the ship in 15-seconds, and the hypersonic missile will hit the ship in 5-seconds.
They are all flying at a randomized flight pattern at ~2G lateral movements. Its not like missiles fly straight at their targets these days, they fly monte-carlo randomly to throw off defense systems.
------------
Now consider the drone swarm flying at 100mph that crossed over the horizon with the 12 cruise missiles. The Drone Swarm will arrive in 2+ minutes. Do your defense systems even care?
Do the drones have any significant amount of lateral G's that they can pull? Can they fly at 2G random flight patterns to throw off air defenses?
In your naval combat scenario, I think you're absolutely correct and the proliferation of autonomous drones will not cause a seismic shift in the battlefield. Unless we go mad scientist and throw in autonomous mobile mines that with a nuclear warhead that lurk around until they home in on the acoustic signature of a CVN's screws from the SCS to Guam. But we already have attack submarines.
----------
I've always imagined autonomous drones as having the most impact in urban combat, or 'traditional' divisional level fights. We've seen the early impact of this technology in Crimea and most lately in the Armenia and Azerbaijan conflict.
And the infrastructure is so much easier - the tomahawk needs a specialist launch vehicle, that is in turn vulnerable to being spotted and attacked. Drones need a backpack and a bluetooth connection (for each 100 drones or so).
Shaped charges are remarkably effective at penetrating armor, penetrating up to 7x their diameter. A 4-inch wide charge will penetrate up to 28 inches of solid steel armor, more than a WW2 battleship's armor.
I don't know how much armor a modern aircraft carrier has, but they're designed for the era of not-very-maneuverable warheads. Send it through the aircraft lifts and blow up the hangar filled with munitions and jet fuel, or punch through the bridge windows and kill the command staff.
Same goes for common tank countermeasures against HEAT rounds, like composite or reactive armor. These are both designed on the assumption that the same exact spot is not going to be hit twice, but you can program a drone to invalidate that assumption pretty easily.
SADARM is smaller than a typical toaster and weighs 12 kg.
An even smaller micro-drone could just attack a tank's gun by setting a small charge or providing blockage that would be devastating when the tank fired.
Pretty sure runways are the reason. VTOL aircraft are too expensive to operate. Helicopters aren't good at attacking ships. Conventional aircraft need a runway: aircraft carriers have to be that big.
Mostly the need to carry big heavy bombs and missiles to attack big targets, and big heavy radars to find those targets, and big fuel tanks to stay up long enough to get to them, which then requires catapults to get them airborne and arrestor wire landing strips to recover them.
This is one piece of possibly good news that it seems like everyone is overlooking.
Automation/Remote control applies not just to drones, but tanks, boats, other aerial vehicles, and maybe troops(after seeing that Boston Dynamics dog robot) etc. Seems like it will be a small squad of highly advanced human troops to move in after the Remotes to secure and maintain control of the area.
So many lives will be spared during wars.
However, this may be unbeatable tech by humans, so any future revolutions will be easily quelled by a gov't with a few drone pilots piloting a drone swarm in a room a thousand miles away.
It'll come down to "how many drones, and of what quality, can your manufacturing base produce?". But this is not that dissimilar from WW2's "how many tanks, and of what quality, can you manufacture?" equation.
Insurgencies will be about getting a small swarm in exactly the right place at the right time.
Or will lead to blending in with civilians, more advanced camouflage, or “hugging the enemy” tactics (staying close enough that they risk harming their own people). Or more terrorism rather than open warfare.
It might be that with these killer drones, we will look back on WW2 style armies in the field as equivalent to 18th century musketeers lining up and firing volleys at each other in bright colored shirts.
When the guns were so inaccurate that getting more shots was more important than actually aiming, that's the natural tactic to employ.
Lining up was hugely important for other members of the army to know what you're doing. If you're 1st in line, you shoot. Back of the line you're loading gunpowder (with everyone else). Middle of the line, you're dropping the bullet into the musket. Etc. etc.
Like a choir performance: its more about having tons of other people doing the same thing you're supposed to be doing. So no one in the army ever "misses" a step in the ~1-minute reload process.
Even if you're not that good at singing, you get "better" in a choir because you just follow the crowd. Similarly, even if you've forgotten a few steps in the long reload process, you just mimic the actions of your neighbors.
-----------
Note that the Redcoat beat the US rifles at pretty much every engagement. It was guerilla warfare that allowed George Washington to survive and last through the war.
Rifles took longer to load. Without precision machining, the bullets needed to be wadded up with paper to stick with the rifling, adding a laborious step in the shooting process. Muskets could fire more shots per minute due to the smooth (though inaccurate) barrels. In army vs army battles, aim doesn't matter, your bullet will hit somebody over there, so there's no advantage to the superior aim that rifles have. You've added 20-seconds to your reload process for almost no tactical advantage at all.
USA experimented with rifles / guerilla warfare, but it really wasn't that successful in the 1700s. (It got the job done but... George Washington lost far more battles than he won). Even in the mid 1800s, the British Square formations were doing decently.
If drones are the weapon of the future, then they will have some kind of advantage over their predecessor. A lot of people in this page are talking as if drones will replace fighters (lol no), cruise missiles (lol no), and other weaponry.
Lets really think about things: What does a Predator Drone really do? How is the military using drones TODAY, and how does that change in the future?
We can see that the #1 advantage of a drone is its __loitering__ capabilities. Missiles can't loiter, and sending a pilot in there to loiter is higher-risk than sending in a drone.
Can __loitering__ help in a fight vs a tank? Yes: if the loitering drone is faster than air support (which it will be, because its at the time/place to support the infantry), then it is reasonable to expect infantry to use a drone rather than call on air-support.
--------------
This is how combat changes. One step at a time, as people incrementally think about the distinct advantages a weapon has over its predecessor. Musket-warfare was such because shooting a bullet-per-minute was better than shooting a bullet-per-1.5 minutes.
Rifle warfare (thanks to the Minié ball) allowed accurate rifles to shoot faster than once-per-minute AND do so accurately. And as such, tactics changed. It was no longer a question of "bullet throughput vs bullet accuracy", an invention happened that allowed both simultaneously.
---------
What advantages do drones have in combat? That's really the big question that needs to be discussed. Today's modern theory of combat is "Combined Arms". We want to invent devices where "Device X" requires special weaponry, while "Device Y" requires OPPOSITE weaponry.
RPGs kill tanks, but are terrible vs infantry. Machine guns kill infantry, but are terrible vs tanks. SAM kill airplanes, but are terrible vs tanks and infantry.
Each "leg" of your combined arms forces the opponent to make a difficult choice. What gun do I pack in my bags today? Where do drones fit in this puzzle?
As far as I'm concerned, drones are looking increasingly like "infantry replacement", more than anything. Their ability to "loiter" in an area, and probably have specialized weaponry (through with a small payload) means you can carry a wide variety of weapons on a wide variety of different drones, in support of a greater overall mission. At least, if future drones are as cheap as everyone seems to be making them sound...
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.
And you're probably underestimating the costs associated with stealth paints / stealth capabilities. Its not easy to avoid modern radar systems. You've pretty much got two choices:
1. Go cheap and swarm -- Ignore stealth entirely, but you pop up on radar like a sore thumb. Which opens up your drones to a variety of automated weaponry. CIWS defense are basically radar-based aimbots that can shoot down supersonic threats like cruise missiles. I find it unlikely that a modern drone would go faster (or cheaper) than a typical rocket lobbed at Israel's Iron Dome, but think of that for a typical defense system you're trying to break these days.
2. Go expensive and stealth -- Stealth capabilities are clearly the future, but their extraordinary costs (F35 project) are well known. Furthermore, stealth capabilities make communication (radio) extremely difficult. If you emit a noise, you can be detected, so a human in the loop (ex: F35) just seems innately more reliable, since they can be "cut off" and enter radio silence for extended periods of time.
--------
Swarming the Israeli Iron Dome with many, many rockets is a strategy that seems to work on occasion. Its not a perfect air defense, so with enough rockets, it begins to miss.
But each of these dumb rockets are launched extremely cheaply: without much computer smarts at all, just flying as fast as possible to make it harder to get shot down. Once we start talking about computer systems (especially in 2021, a year where we've run out of computer chips), your costs escalate severely.
The cheap-and-swarm methodology, especially if you're not really concerned about where you hit exactly (Hamas is mostly doing this as a political / terrorism ploy), is kind of well optimized already. And I'm not entirely sure if an organization like Hamas would switch their rockets to drones in an attempt to beat Israel's Iron Dome. In fact, I'd argue that such a move would be worse than what they're already doing.
Some autonomous weapons blur the line, like the 'sniperbot' recently used to kill an Iranian nuclear scientist as he drove by. The weapon itself was mounted to the back of a truck, and could have lay waiting for days/weeks/months. There is no need for the fighters to blend in, when the weapons can - I imagine in the future where some high ranking military or political figure is sniped by a traffic light (by way of facial recognition or a humble license plate reader)
I'm not fully convinced. The advantage of drones is that they are relatively cheap. That automatically comes with significant range and speed limitations; after all, as someone else mentioned, we already have long-range semi-autonomous drones called missiles. They aren't cheap.
Naval engagements, not to mention aircraft engagements, happen at distances that are just not in the reach of cheap drones. Someone might get a surprise shot at an aircraft carrier in peacetime, but probably not while alert on the open sea. And I'm sure ships will be upgraded with defenses (if they haven't already).
Tanks and infantry, that's a more interesting problem.
They already have CIWS (and a bunch of other stuff) to shoot down incoming supersonic and maneuvering anti-ship missiles. Any drone slower or less difficult to hit than those missiles stands little chance. Drones can extend the eye and ears of a ship but they are not going to be killing ships themselves anytime soon.
If you produce 100 small boats (I'm thinking torpedo-boat size), each of which carry a few thousand cheap drones with a 5km range, then you can do 90% of the job of an aircraft carrier for 10% (or less) of the cost.
https://www.youtube.com/watch?v=1K8LGTwy6Zw
To sum up the conclusion of the several-hour video series: the Aegis Combat System is essentially impenetrable to all but the most overwhelming of forces.
I won't pretend to be some expert on military hardware, I'm a guy interested in flight sims, and that video series just happened to catch my attention. They make a compelling argument though, cruise missiles are one of the methods they try (unsure if supersonic), and they are attacking the carrier group in the absolute worst defensive position it could be in: a narrow gulf surrounded by unfriendly territory on most sides.
Still, not much gets through. If it's not the missile defense system that gets you, it's the F/A-18 patrols. If not them, it's the F-14s that can scramble in minutes. Scary stuff.
Cheap, autonomous, easily-deployed tools - weapons drones, sensors, etc - are a huge change that some are adapting to leverage but few have adapted to respond to.
Which is scary considering the "generations of warfare" model where in fourth generation warfare, the boundary between civilian and combatant is blurred..
And it's fascinating seeing the comments here saying "nah, this won't change anything" when it already is changing things.
And yes, the blurring of the lines is a problem. As others have pointed out, hiding drones in civilian populations is easy.
At the SMS aggregator, it took $10k to get started, $1k/month for a short code, and $0.1/message.
When I discovered Twilio and found it took $0 to get started, $1/month for a long code, and $0.01/message, the world changed.
It became easy enough to build yourself, fast enough to experiment with, and cheap enough to fail. I see the same thing happening here but with lives on the line.
They can still fight the guerilla war among citizens. That's been an incredibly effective strategy because it's so expensive to fight. One might wonder if drones are the key to fighting these wars cheaply. If all you need is a small local base and then a thousand drones to surveil a populace... Well that sounds pretty nightmarish for the citizens, but it would certainly change the nature of war.
And what happens if these autonomous drones aren't programmed to accept surrounder (as seems likely)?
A drone cannot effectively damage a tank or any other major equipment.
The payload a swarm-drone is just way too small. Why have 100 needles when you can have 1 sword.
We have rockets, which are way more effective than any number drones.
A "drone" that can do proper damage is already the size of a plane. Anything less is only useful for recon.
Not useless. The idea seem to be that fighter planes will become something like a command center for the drone swarm. There is value in having a human on site, to take quick decisions for instance, and links to drones can be jammed.
And if you look at the F-35 program, you'll see hints that it is the way we are going. It is a stealth fighter with a heavy emphasis on communication systems. Perfect for managing a fleet of drones. Ok, the execution is far from perfect, but that's another debate.
In the end, it is not entirely unlike the current situation. The difference is that these drones blow themselves up on arrival and are called missiles.
Oh, and swarms of drones won't be cheap, that's the army we are talking about. "Predator" style drones are multi-million dollar machines. A small missile like the Hellfire costs around $100k apiece. Forget your idea of a swarm of hobby quadcopters.
A war like no other?
What is likely to happen is any major nation-state that signs a treaty agreeing to a ban will STILL secretly be researching and building autonomous (or nearly autonomous) weapon systems because they do not want to be left behind.
I suppose nuclear states could in principle rest on their laurels, but at least from a layman’s perspective it doesn’t give an impression of being strategically sound.
But obviously it’s a scary rabbit hole to enter. It’s not fun when game-theoretical considerations or human nature seems to make dubious technological developments inevitable.
It's fairly easy and cheap to make an explosive drone and train a neural network to drive it using off the shelf components and open source software. It follows that it's inevitable that at least semi-autonomous drones will be made to counter radio jamming.
So while the best outcome a treaty could hope for is to prevent mass manufacturing these things by some countries, any motivated small country/individual will be able to craft them.
Regular bombs and guns probably still do most of the killing for quite awhile because of inertia, but I'm pretty firmly of the belief that they are outdated technology or will be soon, especially the regular guns part.
Regular bombs are arguably already outdated by missiles with (frequently autonomous) guidance... but for a lot of applications dumb explosives do work pretty well so those probably stick around for a long time.
But certainly lots of ways to make use of autonomy, and I'm no expert on the best way.
In practice, a drone swarm is an area weapon. It's why it's kamikaze and relies on explosives right now. It can say it has the ability to find a specific person using computer vision and maybe it can, but it definitely can't make sure it only kills that one person and it can't kill from a distance.
No flying platform will be replacing snipers any time soon.
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 think you're overstating the difficulty here. There's three things a drone needs to do:
1. Provide a stable/stabilized platform (for sensors and effectors)
2. Respond to recoil in such a way to make the point-of-impact consistent
3. Not fall out of the sky in response to recoil
I think 1 is demonstrated today with drones flying around 4k/8k movie cameras - a lot of that might be the mount, but you could put an AR-15 in that mount too. 2 is presumably easy as well - just make sure that the drone is in the same physical orientation before each shot as in 'training'. 3 is trivially solved - if your drone is robust enough to withstand the recoil, just let the computer autolevel the drone in whatever new location it's found itself in. If this costs altitude, make sure to fire from farther up.
[1] https://en.wikipedia.org/wiki/Soviet_biological_weapons_prog...
Also I'd really expect HN of all places to understand why banning what is basically computer vision based targeting systems is absurd.
My thought is that ransomware is seeing a resurgence because corporations are increasingly internetworked and hybridized with the global internet, because they have to be. So ransomware has begun to outweigh espionage as a method for extracting money from corporations.
Drones (as used so far, in the remote suicide version) solve a key problem many states have: lack of effective tactical ISR. The US and major powers do not have this problem.
Autonomous drones solve... what problem?
My inclination is that they'll deliver the same value as current-gen ML systems: mass, automated intelligence collection & basic analysis. As well as continuous monitoring and triggering on a pre-configured event (e.g. truck leaves this building, launch missile).
Killing something, once you know what and where it is, is not a problem most militaries have.
Training and paying and allocating large numbers of soldiers to do basic intelligence trawling is absolutely a problem most militaries have.
I think you see this in the US drone approach evolution. Shifting from a single vehicle approach to a survivable system with attritable assets communicating back through stealth communication hubs. Because persistence is the real value.
It would be if there was a real war with a lot of things at once. That statement is only true for peacetime "executive action" wars with a 100:1 force ratio. It is an auxiliary role that has little to do with how things would go if there was a war between somewhat equal powers.
The lessons of enforcing foreign policy on ten people who live in a country with neutral diplomatic status will probably not apply if you change the number ten or the neutral diplomatic status.
Which makes sense, because cruise missiles cost less than radar systems.
And I question whether "build one more drone" is an economically superior answer vs "build one more smart bomb."
This bit NATO forces in Libya[1], for example.
Building drones may well be more scalable and economically viable than smart bombs. Consider a multiuse drone that could be utilized for battlefield recon with its camera, or the user snaps something akin to a BLU-108[2] smart munition onto the bottom of the chassis and now they have a remote controlled smart anti-tank weapon. The drone being multi use could increase the use cases, increasing the order size, decreasing the unit cost.
[1]https://www.washingtonpost.com/world/nato-runs-short-on-some...
> Building drones may well be more scalable and economically viable than smart bombs
That's the heart of the matter. Is it, or isn't it? And for what? The commentary usually glosses over this.
E.g. what's the difference between a JSOW-ER deploying CBU-103s / BLU-108s or a "drone"?
I think the more important future distinctions are going to be between attritable/simple vs expensive/capable platforms, temporary vs persistent platforms, and organic autonomy vs network-reliant. Where some combination of the qualities is optimal for any particular mission.
As an example: Is a mid-air refuelable X-61 without organic sensors but carrying smart munitions a cruise missile or a drone?
For those on the receiving end, I can't imagine its anything more than an academic distinction.
For the forces using such items, it could come down to doctrine, or funding fights. It could be that lawmakers approving budgets will find it more palatable to approve a big purchase of 'drones' instead of 'smart cluster munitions'.
I expect we'll see a profusion of new devices marketed as either, or both, depending on the use case and the audience. It may be that the definition will never really be nailed down.
> I think the more important future distinctions are going to be between attritable/simple vs expensive/capable platforms, temporary vs persistent platforms, and organic autonomy vs network-reliant. Where some combination of the qualities is optimal for any particular mission.
I absolutely agree.
They were acting like moving air-landmines. They key part of this is that they are super cheap to make (cost of a car), and operate, and work autonomously. Thousands of them launched, can wreck havock even on a sophisticated military like the US of Russia.
That's why there is a return of modernized cheaper ww2 type of anti aircraft guns.
China’s New Air Defense System Looks Similar To American ‘Stryker’ SHORAD https://eurasiantimes.com/chinas-new-air-defense-system-look...
That describes Azerbaijan and Armenia's war precisely.
> work autonomously
The TB2 [0] is not capable of autonomous engagement. It's remotely operated for weapons release. And half its key components rely on foreign technology (regardless of what Turkey claims).
The Israeli Harop platform [1] does have autonomous engagement, but intended for the S/DEAD role and is in a different class of technical sophistication and price.
Calling Russian export military equipment "sophisticated" is a bit of a stretch, considering modern delivered models are circa-1995 designs, only recently realized.
Maybe 30 years ago, but they haven't been able to even afford to modernize their own forces, and basically lost 20 years of progress to post-Soviet upheaval and kleptocracy.
We'll see what happens when drones go up against a military keeping pace with the times.
On the contrary, I would expect a site with a usersbase that screams incessantly about "privacy" and has an intense hatred of data-mining dark patterns, dislikes opaque ML models, would be against the lack of privacy and data-mining culminating in an execution by the state.
You'll still see them used in targeted assassinations and some countries will have illegal military stockpiles, but that's very different from these things being treated like regular guns/ammo, being commercially available or legal to build.
Or is the alarm here about the fact that every man and his dog will be able to get their hands on them, when they cost $1000 instead of $1000000?
whereas this "slaughterbot" idea is smaller, low-cost, and with far less top-down control.
At least, as I understand it.
On the flip side, the hardware for Slaughterbots is basically here: TinyWhoops cost ~$100, say ~$100 to upgrade the ARM processor to one that can process realtime video, ~$100 for a shaped charge. So the hardware is super cheap, and once developed, the software to tie it all together has nearly no incremental cost. Once it's out there, that's the genie leaving the bottle.
Maybe the small difference is that cruise missiles generally are not anti-personnel weapons; maybe the article is seeking a ban in the same sense that Ottawa treaty banned anti-personnel mines but allows anti-tank mines. But I think this is something that would need to be explicitly addressed instead of just spouting sky is falling rhetoric.
>‘It seems to be easier for us today to imagine the thoroughgoing deterioration of the earth and of nature than the breakdown of late capitalism; perhaps that is due to some weakness in our imaginations’
https://www.lrb.co.uk/the-paper/v41/n09/jenny-turner/not-no-...
With respect to unconstitutionality, really? Are you also able to acquire and keep on your property a functioning ICBM or nuclear bomb? Arms controls exist in the US despite the 2nd amendment.
They hypothesize reducing the misidentification rate will increase the usage rate, ultimately killing more people (but fewer wedding parties)
This is speculation at best, outright wrong at worst. Prove they're better than humans before employing them.
One soldier made this mistake. What if a faulty image fitting/target classification algorithm deployed in a fleet of autonomous drones was vulnerable to a similar misclassification and started flagging some innocuous objects in people's hands as weapons? Would the rules of engagement be permissive enough for the drones to begin killing those people on their own, would there be human operators in the loop confirming and authorizing the strike?
Any failure of the technology will be scoped away. If the drone blows up a city block targeting a kid with an ice cream cone, the cone will be classified as a potential weapon, therefore tragically the child was a militant.
You will NEVER ever prevent bad guys from creating them by "banning" them.
Look at drugs: all this war / bans / illegality, and the cocaine traffic never ever went down in decades (maybe during covid).
Best course of action is to invest in them to not be left behind
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.
Actually we wouldn't even be able to compete today, things like missiles are already autonomous, and if they weren't we would simply be unable to compete in any battlefield where radio was jammed, and it would give a large force multiplier to adversaries even in battle fields where radio is not jammed.
https://www.droneshield.com/dronegun-tactical
https://www.reddit.com/r/pics/comments/o0kzuq/the_security_o...
I can picture there being something like this with drones.
There are problems to this. The major one being that, as mentioned by the posts on this thread, you have to develop the weapon to really know how to counter it.
A more direct science fiction forerunner of this is Bruce Stirling's _Islands in the Net_ (1988). There is a scene where cheap explosive drone swarms are used in terror attacks.
And of course, there's the famous first line of _Count Zero_.
"They sent a Slamhound on Turner's trail in New Delhi, slotted it to his pheromones and the color of his hair. It caught up with him on a street called Chandni Chauk and came scrambling for his rented BMW - Its core was a kilogram of recrystallized hexogene and flaked TNT."
I am thinking of some other more low tech solutions to whittle them down as well: firing a weighted net; sticky silly string type stuff to tangle and gum up the rotors.
Could they not potentially be better than human controlled devices since they should be able to more accurately target things and create less collateral damage?
Of course, there's always the doomsday terminator scenario where they go haywire and decide to start targeting all humans....
Similar logic was applied to bombers in the Second World War. If bombers could fly into enemy territory, hit key targets behind enemy lines and fly out, then the need for armies would go away. Instead we got The Blitz, the firebombing of Tokyo and Dresden, and the deadliest war in history.
Maybe this time is somehow different, but it's an important historical context to consider. Often the expansion of our destructive capabilities simply encourages more destruction.
Typically the only thing that stops WMD type weapons (not sure if slaughterbots fall into this category) is the enemy having them as well. Assured mutual destruction seems to be the only viable option.
If you give a swarm of drones a kill zone and time limit, then they try them to kill every human in the area, how is that so different from MLRS salvo?
Cruise missiles are already slaughterbots. You just give them location and they kill everyone in that location.
https://www.stm.com.tr/en/kargu-autonomous-tactical-multi-ro...
Surreal.
During the Cold War you started to see drones being used for reconnaissance. The US had the Ryan Model 147 and D-21. Once computers became significantly advanced you started seeing an explosion of different applications. The tomahawk missile is extensively programmable, allowing the user to give it a flight course consisting of a number of waypoints, and digital images along its intended path. It would then compare those images to what it sees on the route and use that for course corrections, ensuring extreme accuracy against targets a thousand miles away in a time before GPS. (fun fact: during the Gulf War, the Navy had to plan their Tomahawk stikes to go through the Southeastern portion of Iraq instead of a more direct route because they could use the mountains and hills for course corrections, while the direct route went over relatively featureless desert). In the late 1980s Israel would employ loitering munitions. Essentially these were suicide drones intended to take out enemy radars. It would loiter above for hours until detecting an enemy radar and then dive. There were also decoy drones. These didn't have a warhead, they were intended to simply fly around acting as bait. Some decoys are fitted with radar reflectors to make them appear larger on radar than they really are, essentially the opposite of stealth aircraft.
Autonomous weapons have been around for a long time and aren't going anywhere. Frankly, I think this is probably better in terms of reducing civilian casualties. I'd rather have soldiers use a precision guided weapon like this than calling in mortars or artillery.
This would be much fairer than the current system of gamified point-and-click murder.
I'm pretty sure any artificial rules which make war "fairer" will just be ignored when real bullets start flying since "it's better to be judged by twelve than to be carried by six" is an actual unofficial rule in the military.