“Fighter Jet Era has Passed”

[sgtmike74]

I’m curious on what some here may think of this story about Elon Musk.

 

https://www.flightglobal.com/fixed-wing/the-fighter-jet-era-has-passed-elon-musk/137017.article

 

The whole thing sounds like we are heading towards Skynet.

 

What do you see has the pros of an autonomous fighter drone?

 

What do you see as the cons?

 

Where does all of this AI leads to in the next 20 years?

[Mars Exulte]

A massive surge in warfare, presumably. Even the more dictatorial countries are politically motivated to limit casualties, lest people become too angry. With autonomous AI, you can operate 24/7/365 for decades on end and nobody will care except the folks getting bombed.

 

There is no ''doomsday Terminator'' scenario, btw. That whole thing is just fodder for bad sci-fi and scare tactics. For such a ''smart man'' Musk is pretty damn dumb sometimes.

[AeriaGloria]

Didn’t he also announce he’s going to make an F-35 competitor...

 

He also said that to a room of fighter jet pilots, while you’re trying to make a competitor. Guy is so emboldened by taking contracts from ULA that he thinks the biggest military project of all time in terms of money is actually within his grasp

 

I think even smart people like Hawking can get wrapped up in Skynet fantasies whic wouldn’t happen. Hawking has his degree in black holes, by AI, so I was pretty surprised when he announced its some monster threat.

 

Everyone is so worried about the control of AI but so few understand AI is just a smart roll, it may do a lot but what is important is how you control it. A bit like a fighter jet actually...

[Birko]

There are a few advantages to kicking the pilot out of fighter jets. They can be smaller and lighter, increasing range and stealth. They can fly to their structural G limits, no longer limited by a mere human being's ability to keep blood in their head. No human error, although AI could certainly have errors, and so could the human designers and engineers and mechanics anyway. Gonna replace them with robots in a few years as well?

 

Edit: if the AI is human designed and has errors, isn't that still technically human error..?

Edited February 29, 2020 by Birko

[MiG21bisFishbedL]

It's just Musky proposing stupid things, again. Much like his hyperloop. This is also absolutely not considering the impact of politics on acquisitions.

 

This isn't the first time this has been said and I doubt it will be the last time. When piloted combat aircraft are a thing of the past, it will be day to be lauded or a day of immense remorse.

Edited February 29, 2020 by MiG21bisFishbedL

[probad]

it's just proxy warfare, we've already long been in it. soldiers are proxies for leaders, bullets are proxies for soldiers, we're just now at the point where proxies get proxies (missiles shooting missiles aka ucavs). ai will be tied up helping us manage our own proxies while making sense of the enemy proxy network for exploitation.

 

the job of the pilot of the future is to direct the exploitation of the enemy network and to that end should be more of a hacker type than an airman. the tactical execution end will be almost completely taken care of by a unified land-air-sea drone complex, all you need are shepherds creatively directing the proverbial finger of god in all sorts of rude and disruptive manners.

 

things won't really feel any different under the thumb of ai, we already live our lives controlled by numbers. just look at how doggedly people follow around sales numbers and other statistics, absolutely lmao

Edited February 29, 2020 by probad
inb4 1.7

[sgtmike74]

Speaking about the ethics. Do you feel it is ethical for a machine to autonomously decide to kill a human, without human intervention? Regardless if that person is an enemy or not?

[probad]

when stat analysts decided what acceptable casualty rates are do you think that's any less impersonal than machine recognition?

 

hot take: ethics is a triggerword used to disable critical thinking

Edited February 29, 2020 by probad

[Birko]

The machine's just a middle man. Instead of a human taking a life, you've got an AI taking a life because a human made it. Even if an AI was 'free thinking' and can make its own decisions, its only making its own decisions because somewhere down the line, a human made it. I don't think it's that different, what about anyone else?

 

Another point is in cases of human error, teaching other humans to not make that same mistake can be hit or miss. If an AI makes an error, potentially a patch to the rest of the AI could prevent that error from ever happening again. That is unless the error is the AI disconnecting itself from whatever network handles them and removing the bit of human control over them.

[probad]

The machine's just a middle man. Instead of a human taking a life, you've got an AI taking a life because a human made it. Even if an AI was 'free thinking' and can make its own decisions, its only making its own decisions because somewhere down the line, a human made it. I don't think it's that different, what about anyone else?

i agree with this, the responsibility is there -- the ai is still owned by someone and so there will always be accountability.

 

theres accountability here on all sides; accountability on the owner of the machine for employing it, and accountability on the opposing side for dealing with it. if you don't want to get shot by a robot then you better figure out how to not get shot by one whether it means using force or diplomacy.

i don't see any issues at all.

[sgtmike74]

How do you all see AI aircraft being implemented? Would be something like the movie “Stealth” with a Uber jet fighter?

 

Would be more like a larger cargo craft deploying waves of smaller drones, each armed with one or two missiles that swarms an area?

[Wizard_03]

I disagree with Elon Musk completely, AI is no where near able to match human decision making at this time. Such a system is fundamentally limited by the information it’s been given and just can’t compete with a humans capacity to reason and make judgments based on new or unknown input. In other words AI cannot improvise well enough yet for even basic combat sorties.

 

The future of autonomous aircraft will be limited to low or no risk scenarios where they are not required to make complex assessments or deal with rapid changes in the mission. Missions such as aerial refueling, decoy, EW, and transportation. Maybe one day we’ll see a robot fighter aircraft, but even then, taking the human pilot out of the equation opens up a whole new can of worms in regards to cyber and electronic warfare. The benefits simply don’t balance with the risks. A human pilot cannot be hacked or jammed.

[Birko]

I do agree that AI might not be quite there yet, but I can imagine that progress is exponential and that AI technology will advance faster and faster.

 

A human pilot cannot be hacked or jammed, but their aircraft's systems can be. At that point a manned aircraft with no sensors, weapons, navigation etc is just as useless as a hacked or jammed AI drone would be maybe? If drones and manned aircraft face the same risk from electronic warfare, is the solution to make manned aircraft with as little technology as possible, or to use AI drones instead as they're crewless and more disposable than a manned jet?

 

I can see drones in the future being propelled electrically and using electronic weapons too whether that be for EW or energy weapons. If the only thing a drone needs is energy- no fuel or missiles or shells- then theoretically you could have 'tanker' drones that are basically flying solar panels go about 'rearming' and 'refueling' other drones to the point where drones can fly almost indefinitely. Interceptions would be faster, and you could have rings of drones flying around your country or around a warzone acting as AEW.

[Mars Exulte]

Speaking about the ethics. Do you feel it is ethical for a machine to autonomously decide to kill a human, without human intervention? Regardless if that person is an enemy or not?

 

Put bluntly there are no ''ethics'' in warfare, as much people like to pretend to civilise it. It's state sanctioned murder for political purposes, full stop. As for whether a human kills a human with a stick, a rifle, or an ''autonomous robot'' it's irrelevant. That one dude's still dead, so the method is irrelevant. Having it or not having it isn't going to prevent or ''humanise'' war. People.been clubbing each other with sticks for millenia and never hesitated. This is no different.

 

Only the political motivation to appease the local populace might change, even then people still can complain about wasted money and foreign deaths, so... probably not that different.

[JUICE-AWG]

I’m curious on what some here may think of this story about Elon Musk.

 

https://www.flightglobal.com/fixed-wing/the-fighter-jet-era-has-passed-elon-musk/137017.article

 

The whole thing sounds like we are heading towards Skynet.

 

What do you see has the pros of an autonomous fighter drone?

 

What do you see as the cons?

 

Where does all of this AI leads to in the next 20 years?

 

 

"When they Pry my F-16C from my COLD, DEAD, HANDS!"

[AeriaGloria]

You might as well just blame the whole military industrial complex, not one country. Each country is responsible for its own military, and declaring an arms race is just one way of feeding that industrial complex.

I’m pretty sure militaries would have developed AI in the same way even in the absence of a big threat, using the justification to prepare for the unexpected.

 

I join the group that thinks fears of AI are overblown, it is just another step in the kill chain, what matters is how people control it. There won’t be some AI uprising where they control the means of everything and production. I actually think it’s pretty irresponsible of people like Hawking and Musk to row fear of AI in the people that see them as a pillar of truth. Extra judicial killings has always happened, it’s just obvious now

[Emu]

Speaking about the ethics. Do you feel it is ethical for a machine to autonomously decide to kill a human, without human intervention? Regardless if that person is an enemy or not?

I think the real problem is that waging war will require no buy-in from human military personal, and governments will be able to wage any kind of war they like, even against their own people. That said, what can you do if the opposition is doing it?

 

I don't know if I've gone political here, but it's kind of political, however it should be okay as long as it remains non-nationally biased or partisan.

Edited March 1, 2020 by Emu

[Pikey]

'Somewhat' agree.

For basic interception duties, peacetime or full convential war, a drone has the "potential" to be faster to the point of the intercept and accept a greater amount of risk than human equipped interceptors, but I'm not so sure about 'fully autonomous', ever. There is so much issue around detection and identification, MITL comms being degraded or unavailable, ECM heavy environments, and state level interference with code and chip technology that I can't see a way to guarantee armed drones be allowed to carry out any life threatening persecution of threats with deadly weapons. Maybe the US could sign that off eventually but it's a long way off to say, hey we are giving this machine the ability to kill people by itself AND it can operate without a man in the loop, fully autonomously.

 

AI is still written by humans, it still fundamentally has a human at it's heart wether it is realtime or not. Whether we scapegoat a developer or a soldier with a rifle, a general or a politician, blame eventually lands on someone human, AI is simply a projection of human intention, but written in a flawed way.

But there can be a use, some scenarios can accept it. I would say any scenario where the human makes the decision to use lethal force and then the machine carries it out, is already abundantly used and "autonomous", in this case, can be extended further and further afield. We already have vast distances across the globe where this has been done for many many years, so much of this is in the wording.

Specifically on fighter jets, I have some doubt. We operate gliders and very small manned vehicles all across the globe. Some of these without electronics. People fly where they shouldn't all the time, wandering across the superbowl, losing comms, no transponder. Visual recognition is not strong enough to be infallible. Without a way to identify a target, autonomy is useless.

[Quadg]

Put bluntly there are no ''ethics'' in warfare, as much people like to pretend to civilise it. It's state sanctioned murder for political purposes, full stop. As for whether a human kills a human with a stick, a rifle, or an ''autonomous robot'' it's irrelevant. That one dude's still dead, so the method is irrelevant. Having it or not having it isn't going to prevent or ''humanise'' war. People.been clubbing each other with sticks for millenia and never hesitated. This is no different.

 

Only the political motivation to appease the local populace might change, even then people still can complain about wasted money and foreign deaths, so... probably not that different.

 

why build war machines when you can build death factories?

 

more efficient..

 

and very nazi :)

 

war is politics by other means.

[Toxic]

Anyone not surprised by this? It was said by someone who is in the business of automation. You keyboard/Wikipedia warriors should know that.

[Fri13]

What do you see has the pros of an autonomous fighter drone?

 

Aircraft body limited G force capabilities. You can make a drone pull as much and as long you want without limitations.

 

Great for "one-time-trip" missions, where you don't anymore need to care does the aircraft come back or not as you don't need to deal with pilot dying or being captured etc. You just blow the done after high risk mission completed.

 

What do you see as the cons?

 

AI is not at all so advanced that it can perform a autonomic decisions. You need to babysit it and make decisions for it.

 

You can't put the drones flying in the situations that are dynamic. That is again the main reasons why A-10C is wanted to keep in flying, because there is always a soldier in the cockpit that has different perspective to the battle and can react alone to new threats and situations with big firepower, without going through some chain of command.

 

 

No, I don't trust anything to datalinks but only consider case where you need to have a autonomic operations capability. Flight with datalinks and mission updates but aircraft should be able fly without anyone sitting in control.

[Bozon]

There is a difference between autonomous and unmanned. Unmanned is a misleading term because there usually a man in the loop. The detailed operation of the vehicle is done by the vehicle (navigation, flight control, etc) but the orders come from a human operator.

 

At this level already man manned air operations become obsolete. Attack helicopters are the 1st to go. Armed drones can replace them for 90% of what they do and do it just as well if not better. Then of course you have the various aerial recon missions - a man on the ground is good enough, not need to put him on the plane. The more complicated the mission the more you still need the carbon unit in the cockpit. This is why deep strikes still need manned jets.

 

In “manned from afar” unmanned vehicles the uplink/downlink is... well... the weakest link (excuse the pun). This is why there is a push for autonomous machines to ensure the execution of the mission even if the communication with the human controller is severed. Swarms of fully autonomous vehicles that will saturate all defenses are still far away in the future, but taking the human out of the semi-classic vehicle and putting him safe on the ground is already here.

[pctoptech]

Machines cannot make decisions on right and wrong. When not to fire even though the orders say too. Nor can they make those uncalculated moves that only humans can do, that "Gut feeling".

[Pikey]

There is a difference between autonomous and unmanned. Unmanned is a misleading term because there usually a man in the loop. The detailed operation of the vehicle is done by the vehicle (navigation, flight control, etc) but the orders come from a human operator.

 

At this level already man manned air operations become obsolete. Attack helicopters are the 1st to go. Armed drones can replace them for 90% of what they do and do it just as well if not better. Then of course you have the various aerial recon missions - a man on the ground is good enough, not need to put him on the plane. The more complicated the mission the more you still need the carbon unit in the cockpit. This is why deep strikes still need manned jets.

 

In “manned from afar” unmanned vehicles the uplink/downlink is... well... the weakest link (excuse the pun). This is why there is a push for autonomous machines to ensure the execution of the mission even if the communication with the human controller is severed. Swarms of fully autonomous vehicles that will saturate all defenses are still far away in the future, but taking the human out of the semi-classic vehicle and putting him safe on the ground is already here.

 

 

Wording... but, we've been doing unmanned deepstrikes for 20 years +. What do you think a cruise missile is? :)

 

Expanding unmanned units to include more helicopters, interceptors and things we do not already do it absolutley on the agenda, but the decision process is not in the realms of being taken away from battlefileld commanders any time soon, and that decision process, or AI is what I believe Musk was inferencing

 

We will see more of the same, but the day a script chooses to automate killing humans without a human in the loop, is the day I ask the planet to stop... because I'll be getting off.

[SmirkingGerbil]

Coming from a guy, that unleashed on the world, the ugliest truck ever known to mankind . . . meh.

 

 

I designed that truck when I was eight years old. It was a shoebox that I cut down, and taped together to create that exact profile. I then took the wheels off my Tonka Truck, and shoved the metal axles through the cardboard, and stuck the wheels on. Looks exactly like the Tesla Cyber Truck, from the 60's.

 

So I rest my case, unleashing a truck, designed by an 8 year old, and then pontificating on AI, does nothing for credibility. Meanwhile, Rivian has the ability to do a 360 in place, and looks a damn sight better.