There is a common myth being propagated by senior civilian and military leaders that unmanned assets are some kind of magic assets that increase our military capabilities to undreamed of levels.
For example, here’s a statement by Michèle Flournoy, former Under Secretary of Defense for Policy and a top contender for the job of defense secretary in a possible Biden administration,
“We want to augment our manned forces with unmanned systems that are still controlled by a human being, but that dramatically improve ... our ability to project power …” (1)
Ms. Flournoy suffers from the common misconception that unmanned equates to greater capabilities. This is bilgewater.
As a general statement, unmanned systems have no greater capabilities than equivalent manned systems. For example, the Large Unmanned Surface Vehicle (LUSV) that the Navy is currently trying to acquire will not be faster, more stealthy, or carry more missiles than an equivalent manned vessel – it simply won’t have crew on board. The Fire Scout unmanned helo is not faster, stealthier, or able to carry a heavier payload than an equivalent manned helo – it simply doesn’t have crew on board. The MQ-25 Stingray will carry no more fuel and have no greater range or endurance than an equivalent manned aircraft – it just won’t have crew on board. And so on.
Unmanned assets don’t have any magical capabilities – they just don’t have crew. They’re built using the exact same components and equipment as manned assets so why would there be any huge improvement? There wouldn’t! Yes, there could be some scattered marginal improvements in performance such as an unmanned aircraft being able to tolerate greater g-forces than an equivalent manned asset. However, there are also significant decreases in performance from not having an intelligent human controlling the asset. Even remote controlled unmanned assets which have a human in the loop have significant performance decreases due to the communication lag time and decreased situational awareness for the remote operator.
Okay, so there’s a tendency to overestimate the capabilities of unmanned assets – what’s the big deal? The big deal is that this overhyped belief is being used to determine our future force structure and remake our armed forces. The Navy, for example, is going to decrease the number of surface warships (Burkes) in favor of unmanned vessels despite the unmanned vessels being markedly inferior in capability. The justification is that unmanned vessels will somehow, in some unexplained way, provide greater overall capability than the same number of manned Burkes. That’s right … Individually inferior unmanned vessels will somehow prove superior in the aggregate than the same group of individually superior manned vessels. Does that seem plausible? Of course not! It’s insane.
Let’s change gears, for a moment, and consider the Army’s quest for an unmanned replacement for the Bradley Infantry Fighting Vehicle (IFV). The replacement won’t have any engine, gun, munition, speed, or range that an equivalent manned vehicle wouldn’t have and yet the Army is spending gazillions of dollars to achieve the unmanned aspect of this. Sure, I get that in combat you’ll save a few lives but you won’t gain any combat performance and, almost assuredly, the unmanned replacement will perform worse than the manned equivalent because our feeble attempts at artificial intelligence are nowhere near ‘Terminator’ levels of sentient intelligence. There is every reason to believe that the unmanned replacement IFV will suffer much higher loss rates than manned vehicles. Now, if the unmanned version cost a small fraction of the manned version so that we could build several for the price of one manned vehicle and we could, literally, throw waves of vehicles at the enemy then it would be worth it but since the vehicle will be identical to a manned equivalent in every way except seats for the crew, the cost will be the same. In fact, the cost may well be greater since the remote, automated control costs additional money.
So, why the big push across the military to develop unmanned assets? Our limited combat experience with unmanned aerial vehicles (UAV) demonstrates that UAVs are nothing but cannon fodder in a contested environment. Yes, there’s a small niche for unmanned assets in the highest risk missions so as not to endanger operating crew but that doesn’t justify the kind of wholesale conversion to unmanned assets that we’re seeing the Navy push for.
The unmanned push is yet another example of latching on to an idea and leaping right over the justification and straight into implementation. We did this for ‘littoral’ and the Navy got the LCS. We did this for minimal manning and the Navy got a maintenance and readiness hole that we have yet to dig ourselves out from. We did this for gender equality and we got a Marine Corps that is demonstrably less capable. We did this for long range naval guns and the Navy got the Zumwalt with no gun.
Unmanned offers no substantial increase in performance and yet we’re converting our entire military to it. We desperately need to pump the brakes and establish actual performance gains, if any, that are sufficient to justify this kind of wholesale makeover of our armed forces. In the Navy’s case, the makeover is trading capable surface ships, the Burkes, for small, marginally capable unmanned vessels. This is a serious mistake.
______________________________________
(1)Defense News website, “Flournoy: Next defense secretary needs ‘big bets’ to boost ’eroding’ deterrence”, Aaron Mehta, 10-Aug-2020,
I love that you mention the Bradley here to demonstrate the folly of unmanned. I think the biggest lesson that one could learn is the fact that the Bradley hasn't been replaced yet. 3 major programs in (Future Combat Systems, Ground Combat Vehicle and OMFV) and 2 lost decades with nothing to shown for. What's the point of these technology if it never trickle down to the average unit in the field? We are chasing dreams and vision while the Chinese is incrementally surpassing us in every field.
ReplyDeleteThe part about the unmanned benefit, I get it. The reason why it is so overhyped is because it presumably poses no risk to the operator. No risk means nobody getting home with a scratch and burnt and that's what most young people care about nowadays. Is it capable? No one cares anyway. Does it complete the objective? Who cares? And just like that, combat effectiveness get pushed to the bottom of the list and inclusiveness and political correctness is promoted as heroism.
That's good enough for the "terrorists" wars we have been fighting but a peer war would be our doom. . .
It's the Endless Quest to Reduce Manning, whose reason is not entirely clear yet.
ReplyDeleteMaybe there's just not enough worthy men willing to serve in today's armed forces?
Wouldn't surprise me.
Lonfo, the Robot Union has yet to ask for retirement & repair benefits from the Navy, once they do that the Navy will lose interest.
ReplyDeleteThat's a good point, of course.
DeleteAnd politically the benefits are clear: nobody gives a damn if Unit B888XX41R-9 blows up in Southern Whateverstan where US forces are for dubious reasons, while body bags can make the public angry and demanding for accountability.
There's just the little detail of the entire concept not working against anyone but goat herders, but I guess that's no big deal.
Let's just hope China never figures out how to hack those robots, then?
Not that the navy doesn't have its own issues, but it is often also constrained by the hand it is dealt. Personnel costs of the all volunteer force go up over time, especially when Congress adds further incentive to leave for a better paying commercial job that keeps people in one place with their kids. On the other end, you have less and less people who qualify to serve. Congress and the people fix that, not the navy. Should acquisition be reformed so the service provides sound requirements and makes sound reviews of proposed solutions, you bet. Should we look more to history in understanding how to bring new technology into the fold, yep. I mean we used to have sail and steam propulsion at the same time. We once blew up a Secretary of the Navy. We need people going Myth Busters on problems to make fast, practical solutions. If you can't solve an engineering critical thinking problem in peace time, you certainly aren't going to out think an opponents strategy in wartime.
ReplyDeleteI think you just exposed the elephant in the unmanned room...!!!
ReplyDelete"Unmanned=same or lesser capability at greater cost."
⬆⬆⬆ This is somthing that needs to be understood and until it can be decisively overturned, should bring the unmanned push to a halt.
Its also a great line for a "Change My Mind" meme...
Spot on, as usual. A couple comments:
ReplyDelete1) Part of the math of the 'savings' is not needing crews, but, as you have pointed out, the support required for unmanned vehicles actually means paying for more people, not less. It's just accounted for differently.
2) Thirty years ago, I might have said "Give them the benefit of the doubt. There are things they know that they can't tell us, and they get it right more often than not." Today, we have to start with the assumption that they are about to get it completely wrong. Because that's what they have proven to do time and again. It would be more surprising if they said "we are pursuing cutting-edge R&D while building systems based on the latest proven technology as demonstrated through working prototypes."
How is this not a hot political topic nationally? What's the right dramatic headline needed to cut through the noise and get people's attention: "After 25 years of procurement debacles, the Navy doubles-down are becoming an experimental-only, non-functional 'floating IT' branch of the armed forces with minimal combat power. They are considering the new motto 'netboats not gunboats,' but aren't sure it adequately communicates the complete lack of reliability of their designs."
I believe Russia tried out some unmanned tanks in Syria. They all failed due to mechanical issues and especially communication ones.
ReplyDeleteI agree. This is just another military contractor sales scam. The question I ask is why jump from two pilots per aircraft to none? Does an H-60 require a co-pilot given that AI can double check the pilot and land the aircraft if needed. AI replaced navigators 20 years ago, maybe we can shed co-pilots on smaller aircraft. Let's try that before we consider unmanned. If a co-pilot is still needed, that really kills the unmanned concept. Such logic leaves the unmanned salesmen stunned.
ReplyDelete"The question I ask is why jump from two pilots per aircraft to none?"
DeleteThere are many nuances to that question and I'm not sure exactly what aspect you're considering. I'll offer some thoughts that may or may not be relevant for you.
You note that AI replaced navigators and you're correct - within limits - and it is those limits that form the core of the unmanned discussion. Yes, an "AI" (the quotes are intentional to indicate a pseudo-AI) can navigate in the sense that it can verify gear down or set a course to the next waypoint, and so on. What it can't do is decide what to do when the mission has gone to hell. It can't make judgement decisions about whether to risk the aircraft to try to refuel that friendly plane that is in trouble in enemy airspace or save itself for the greater good. And so on.
Our AI is not at the level of human judgement. Sorry for belaboring the obvious. So, with that limit firmly in mind, why are we pushing so hard for unmanned assets? There is no overall personnel decrease (and likely is an overall increase according to the Air Force). There is no cost savings. There is a performance decrease in most aspects. So why the push? In the Navy's case, I think it's a budget move to try to redirect manning money (operational costs) to new construction. I've stated many times that designing a force structure to the business case is idiotic. You design forces to the combat effectiveness case.
I'm pretty sure you're not discussing the co-pilot scenario specifically but, instead, using that to illustrate your larger point. I'll offer one thought, specifically, on the co-pilot scenario that may hold some value for the larger issues. A co-pilot's main value, aside from sharing the workload and supporting the pilot is to act as a redundant backup system FOR HUMAN JUDGEMENT. If the pilot is incapacitated in combat (or for any other reason) the co-pilot can take over the judgement function and continue to operate the aircraft with human judgement still in the loop and, therefore, combat-value. It's somewhat akin to a backup radar. It also illustrates the flaw in the Navy's minimal manning push. In combat, people will die and without 'redundant' people to back them up, we'll wind up mission killed due to very few casualties. But, I digress ...
The lesson from this is that if we commit to unmanned assets with extremely limited "AI" and we lose the ability to insert our judgement into that "AI" then we've lost the entire asset because the "AI" has no ability to make its own judgements and continue to operate EFFECTIVELY. Yes, there will be some kind of default 'return to base' or 'orbit and wait' or some such programmed command for when we lose touch but how does that help us in war? It doesn't!
Your suggestion about easing into unmanned by substituting "AI" for certain tasks and positions is the only approach that makes sense. It's doing the hard work, one small step at a time, instead of reaching for fantasy levels and, inevitably, failing.
I don't know if this rambling bit addressed any of your original point or not so take it for what it's worth. Feel free to expand on your thought or set me straight if I've totally missed the mark!
Maybe some astute reader needs to do a post on the proper use of AI and the proper path to get there? Just saying … :)
"...our feeble attempts at artificial intelligence are nowhere near ‘Terminator’ levels of sentient intelligence..."
ReplyDeleteThis ^
We should be spending our AI research and development resources on artificial intelligence that helps sort through mountains of electronic information overload for our human sailors.
For example:
the ability to sort through the clutter for an Aegis operator so that they can fully utilize the system,
or identify targets and prioritize them for an attack aircraft pilot,
or help a sonar operator recognize sonar anomalies to find enemy subs,
and the list can go on and on...
Lutefisk
Most of your suggestions already exist. Aegis has a full auto mode that prioritizes targets and acts on its own. Our sonar systems perform automated analyses to locate subs. And so on.
DeleteIronically, our insistence on maintaining a man in the loop is what led to the Vincennes disaster! Aegis had the airliner correctly identified as a non-threat but the human operators misinterpreted the data and acted incorrectly.
What our automated systems can't do is apply intuition, tactical knowledge, experience, and the like. For example, our sonar systems can't anticipate enemy submarine actions and know where to look.
We've already taken the steps you suggest. Now, we have to learn to balance the human and machine inputs correctly and learn to recognize the limits of each.
Sorry, you can tell I'm not a navy person.
ReplyDeleteI guess I just look at the unmanned things and it seems like a preposterous set of ideas...there must be some other way that those resources can be applied to enhance actual fighting capabilities.
Lutefisk
Without repeating many posts, the heart of the matter is that unmanned platforms are invariably LESS CAPABLE than manned ones. So, our push to more unmanned platforms is, quite literally, a push to less capable forces! Our professional warrior leaders know this and yet they're doing it anyway. This is baffling.
Delete"Sorry, you can tell I'm not a navy person.
DeleteI guess I just look at the unmanned things and it seems like a preposterous set of ideas."
Too bad you're not a Navy person. We need that kind of thinking, all the way up to CNO (the "official" one, not the ComNavOps one, unfortunately).
Thanks CDR, I bumbled onto this blog a ways back because I saw an article about modernizing battleships.
DeleteI've learned so much from CNO and some of the sharp posters on here.
I wonder if there are any active navy folks that post here?
I can see using small UAVs, USVs, and UUVs for ISRT (intel/surveillance/recon/targeting) missions. Past that, I just don't see it.
ReplyDeleteMy cynical side says the reason we want drones is because you don't have to write letters to their mothers when somebody blows one up. My problem with that is that the more we make war antiseptic, the greater the temptation to barge in and fight no-win wars over stuff tat doesn't really matter to us.
There is a currently ongoing effort to discredit Robert E. Lee, but I think he nailed it when e said, "It is well that war is so terrible, otherwise we should grow too fond of it."
" the reason we want drones is because you don't have to write letters to their mothers when somebody blows one up"
DeleteAside from the Navy's desire to transfer personnel (operating costs) to new construction, we're also dealing with a societal influence that makes us pursue unmanned assets. We've become lazy as a society. We no longer are willing to do the hard, dirty, unglamorous work that accompanies real progress. Thus, our naval leaders - products of that society - continually search for the quick, easy, technological leap instead of putting in the hard, grinding, incremental work to advance our capabilities. Every problem we face, we look for quick technology fixes instead of hard work solutions.
Recall the IED problem in Iraq? Instead of applying better tactics, driving off road, implementing secure routes, and improving training, what did we do? We spent hundreds of millions of dollars pursuing technological solutions … none of which worked and all the while we continued to let ourselves get blown up. Every vehicle in the Army is designed for off road capability so why did we continue to drive on roads where the IEDs were???? Get off the roads, use different routes, get unpredictable, and shoot suspicious people. Sure, that's dirty, grimy, slower, and less efficient but it solves the problem completely and doesn't require expensive new technology which didn't work anyway.
US military are repeating "stealth" progress with unmanned and AI, at first LO was top secret small programs. Then once it went public and was seen as a silver bullet after GW1, they decided in their infinite wisdom to dump it on everything even when it isn't needed, not worth it or should be understood that LO is just another element in the equation of a weapon system. Same seems to be going on here: UAVs,USVs,etc with AI have undoubtedly a role to play: very small robots and crafts for the Army should be useful, Predators and Reapers and the like for long and boring over permissive countries are fine, high risk high value ISR can make some sense AND that's really already a lot!
ReplyDeleteBut we're seeing it already, DoD wants to push unmanned everywhere to do everything and its just another waste of money. The danger this time compared to LO is that pushing unmanned everywhere really endangers the remaining manned forces when all these fancy gizmos get crushed by the enemy.
Manned or unmanned weapons have their own roles. Saying one should totally replace the other (or the other way round) is not right. For instance, UAV is best tools for mine sweeping, which is very danger to manned ships.
ReplyDeleteNavy without resource problem and good management should pursue both. Current UAV problems mainly come from insufficient technologies on them. Key is to conduct R&D.
As technology progress, unless Navy progresses ahead of any potential enemies, even allies, there is no way that US can continue dominant in the world.
Sad thing is that fewer and fewer bright students choose STEM but want to be lawyers.
Right now on the air side it appears almost all autonomous vehicles are either glorified remote control A/C, or missiles.
ReplyDeleteThe design concept they should be looking at is vehicles able to do a single task well. Don't try to design the vehicle. To misquote Clinton, "It's the software stupid." Take an existing design, like say the F-16. Turn it into your flying computer (QF-16). Then tinker with the software to your hearts content. When you get something useful, then give folks a yell.
After the software is validated, design the task specific vehicle. Where you get a little savings here is that you don't have to have more than you need for the mission.
At the end of the process you a useful missile toter (tanker, jammer, etc...). You break it out of storage when needed. You don't need to constantly fly it in order maintain its skills. Throw it in storage till you need it.
This approach doesn't get miracle do everything wonder machines, but it can get you basic useful tools that you can afford to have and use up in war.
This, this I agree and support if the Navy is making it the forefront of AI revolution. On the other hand, what value does a limited F-16 tanker or jammer would o for us in exchange for a perfectly capable second-line fighter?
DeleteYou have to remember when the dust settles and the bombs dropped, these storage aircraft is gonna be only thing we have to reinforce any front-line units almost immediately. Converting them into "missile toters" (did you mean loiters?) and second rate non-combat-capable aircraft is a risk that maybe not worth it at the end.
However, I think there is a balance that can be found and we just need to tinker and experiment to know what it is. Maybe using air-frames that only have a few flight hours left?
On other notes, the Air Force back in 2017 experimented with the idea of having a strike UCAV using a QF-16. Maybe, they are following what you are proposing. Attached link for reference:
https://www.popularmechanics.com/military/research/a26028/f-16-drone-have-raider-ii/
Unmanned devices can achieve great tasks than manned, for instance:
ReplyDelete1. Mine sweep
2. Front line reconnaissance
3. Cruise missiles are actually UAVs
4. Extend ranges deep into enemy's territories (human could be nervous)
Real key difficult is remote control under hostile conditions. Under normal conditions, you can control them as long as you can send/receive EM waves to them. Key problem is signal strength attenuate with distance nonlinearly. While far from you but close to enemy, nations with good electronic warfare capabilities can either make your devices dysfunctions or destroyed.