Be it known that we, JOHN M. BROWNING and Marrana' S. Browning, of Ogden, in the county of Weber and Territory of Utah, have invented a new Improvement in Magazine Guns; and we do hereby declare the following [...] an improvement in the construction of guns whereby the firing of the gun after the first discharge may be made automatic so long as cartridges shall be supplied. [1892]
What's the point in the Army developing its own, parallel autonomous vehicle technology? Why not just have DARPA funnel funds to Google or someone and piggyback off their research?
Because the Google car works with pre-registered and mapped out terrain, and the enemy rarely will let you pause for a few months and run a bunch of ladars through their area of control? Because the google car is trying to solve the problem for roads and not figure out whether this stream is fordable? Because the google car's autonomy has never once had to worry about silhouetting on the top of a ridge line?
I'm 95% certain that each one of the vehicles there was produced by contractors for a government agency (and if these particular vehicles weren't, certainly such contracts do exist- I used to work on one). If Google wanted to compete for a contract it certainly could, but its experience isn't particularly helpful in attacking the problem set that the Army is interested in.
Future of warfare is going to something more like Command And Conquer game, where operators will direct autonomous machine with a click of a mouse and warfare will be mostly done at a distance....
Future of symmetric warfare, sure. Future of asymmetric warfare looks a whole lot like Skynet sending robots to kill humans, but Skynet takes a mandated 15 minute coffee break every two hours.
The ethics of asymmetrical warfare with one side nearly completely autonomous seem like very murky water.
Imagine Iraq occupied with pretty much only machines and Iraqis never actually see an American. And if this removes the threat of death from one side, what does that do to the cost of war? Could it actually make it more common because it is less politically risky? It feels inevitable, but it also scares me.
It would certainly make it more common. The cost in American soldiers is the main thing that slows down the US from starting minor wars: the population are not tolerant of losing soldiers.
Other countries are more willing to lose soldiers but refrain from war for other reasons: the economic cost, fear of losing, loss of standing in the international community, or simply having no enemy they particularly want to fight.
Also, why just other countries? Asymmetrical (class) warfare is very popular, here, now. I don't see the kind of people waging it having moral setbacks about other kinds.
Civilian casualties have been a big part of war. It seems too soon to say an automated army won't have the potential to massively increase civilian casualties given that it lowers the cost to the attacker and robot don't care if they engage in atrocities, etc.
You would think that, but if the advent of drones is anything to go by, it's going to simply make things worse.
When you have to put your own humans in a combat situation, you think twice before sending your men to their potential doom. With robots and unmanned vehicles, you can just keep throwing them at the problem, cost is the only concern.
A bloodless war is unable to inflict the necessary pain to force people to change their minds. The actual military logic goes like this...
If the source of a country's military might is the ability of the civilian population to produce goods needed for the war effort, then the civilian population becomes a legitimate target.
The result of that logic in WW II was mass bombing campaigns which were capable of "dehousing" 100k+ people in a single night.
Operators? Machine learning has already come up with the ultimate zerg rush. Who needs operators? We could just flip coins but they would probably argue over the denomination.
There was a fairly impressive display from Russian manufacturers of all kinds of autonomous mobile fighting units and fixed sentry emplacements recently. So it appears not building them is no longer an option. Some of them were even new.
This seems like an obvious step - why put a soldier or soldiers at risk to scout around the corner or breech a building when you can send in an armored quad?
Not only that. Friend-or-foe identification can boil down to "only targets shooting at me are hostile". That policy isn't any good for human soldiers but robots operating on that principle could avoid a lot of collateral damage.
> Friend-or-foe identification can boil down to "only targets shooting at me are hostile".
And those that shoot at allies. And those that haven't shot yet, but are aiming in the wrong direction. And those that are just reloading their mortar. And those that are setting up an explosive trap for the robot.
There are a lot of ways to be hostile that go beyond firing a weapon directly at the robot, and if it can't handle those, it will be exploited by humans pretending to be harmless.
Or just any humans in the area. Since our forces are all robots every human remaining must be in support of the enemy. Civilians should have fled as refugees. (Refugees will be denied entry to the US, on the assumption that they're terrorists.)
I'd say that's sarcastic, but I sadly expect such reasoning from much of the US military leadership these days.
One step closer to the utopia dictators and despots have dreamed about for millennia. No longer will they have to depend on the loyalty of Praetorians or armies to protect them and put down the unruly peasant and proles. Soon they will have legions of obedient killer robots to ensure their rule.
The enormously high cost of war (both economical and humanitarian) is the main reason why our time is the most peaceful in human history. Do not change that.
Little known fact: one of the first self driving cars was a 'Humvee' (real name HMMWV) which was equipped by DARPA with cameras and could navigate a road with no other vehicles, and go straight through a 4 way intersection after stopping (again, no other vehicles). This was in the mid 1990s. Would love a link to this project if anyone has it.
As I recall this was out of CMU. I also recall that they got smoked by Stanford in the actual Darpa challenges, which was slightly surprising given their head start.
I seem to recall Stanford used a far less intimidating VW SUV instead of a giant Humvee.
This is far from being deployed, per another article on the same demo (a much better article, if you are interested in learning more):
Much of the technology is there to drive robotics and autonomy into maneuver formations, but when it comes to developing the tactics, techniques and procedures, the Army is figuring out “how we want to massage this,” said Robert Sadowski, robotics chief with the Army‘s Tank Automotive Research, Development and Engineering Center. “The next 10 to 15 years will help us figure out how we want to embed robotics and autonomous systems into the formation.”
52 comments
[ 3.8 ms ] story [ 105 ms ] threadYeah, how exactly can that be automated?
> Yeah, how exactly can that be automated?
Be it known that we, JOHN M. BROWNING and Marrana' S. Browning, of Ogden, in the county of Weber and Territory of Utah, have invented a new Improvement in Magazine Guns; and we do hereby declare the following [...] an improvement in the construction of guns whereby the firing of the gun after the first discharge may be made automatic so long as cartridges shall be supplied. [1892]
https://www.google.com/patents/US471782
Also, combat zones != California suburbs.
We're just going to leave that one tee'd up, aren't we?
I'm 95% certain that each one of the vehicles there was produced by contractors for a government agency (and if these particular vehicles weren't, certainly such contracts do exist- I used to work on one). If Google wanted to compete for a contract it certainly could, but its experience isn't particularly helpful in attacking the problem set that the Army is interested in.
Probably with less casualties overall...
Oh wait, it's the present.
Imagine Iraq occupied with pretty much only machines and Iraqis never actually see an American. And if this removes the threat of death from one side, what does that do to the cost of war? Could it actually make it more common because it is less politically risky? It feels inevitable, but it also scares me.
Other countries are more willing to lose soldiers but refrain from war for other reasons: the economic cost, fear of losing, loss of standing in the international community, or simply having no enemy they particularly want to fight.
Also, why just other countries? Asymmetrical (class) warfare is very popular, here, now. I don't see the kind of people waging it having moral setbacks about other kinds.
Civilian casualties have been a big part of war. It seems too soon to say an automated army won't have the potential to massively increase civilian casualties given that it lowers the cost to the attacker and robot don't care if they engage in atrocities, etc.
When you have to put your own humans in a combat situation, you think twice before sending your men to their potential doom. With robots and unmanned vehicles, you can just keep throwing them at the problem, cost is the only concern.
If the source of a country's military might is the ability of the civilian population to produce goods needed for the war effort, then the civilian population becomes a legitimate target.
The result of that logic in WW II was mass bombing campaigns which were capable of "dehousing" 100k+ people in a single night.
See https://www.youtube.com/watch?v=tPnppCelvk0 for an hour long documentary on how our world came to accept such logic as reasonable.
There was a fairly impressive display from Russian manufacturers of all kinds of autonomous mobile fighting units and fixed sentry emplacements recently. So it appears not building them is no longer an option. Some of them were even new.
Or have I been reading too much SciFi?
And those that shoot at allies. And those that haven't shot yet, but are aiming in the wrong direction. And those that are just reloading their mortar. And those that are setting up an explosive trap for the robot.
There are a lot of ways to be hostile that go beyond firing a weapon directly at the robot, and if it can't handle those, it will be exploited by humans pretending to be harmless.
I'd say that's sarcastic, but I sadly expect such reasoning from much of the US military leadership these days.
That reminds me of a classic scene from the original Robocop:
https://www.youtube.com/watch?v=Hzlt7IbTp6M
That which controls the wrench controls fate.
/seriousness
Much better to focus on science, and technology and grow peacefully.
I seem to recall Stanford used a far less intimidating VW SUV instead of a giant Humvee.
http://velodynelidar.com/blog/it-began-with-a-race/
Much of the technology is there to drive robotics and autonomy into maneuver formations, but when it comes to developing the tactics, techniques and procedures, the Army is figuring out “how we want to massage this,” said Robert Sadowski, robotics chief with the Army‘s Tank Automotive Research, Development and Engineering Center. “The next 10 to 15 years will help us figure out how we want to embed robotics and autonomous systems into the formation.”
http://www.defensenews.com/land/2017/08/25/us-army-tackles-t...