There’s one major element for AI use in warfare you skipped.
Every modern war has been a contest of industrial capacity. More and higher end weaponry has proven to give a decisive advantage. But the more complex modern weapons are extraordinarily expensive to build.
Part of the reason for the high cost is a modern missile, torpedo, submarine, jet fighter, and so on uses far more components than past weapons, with many complex embedded electronic systems and many new ways for the weapon to be unavailable from an internal failure. Materials are also much more complex and difficult to build, with higher precision required as well. For example, compare carbon fiber to cloth or riveted aluminum sheets for aircraft wings. Higher speeds for the aircraft required tighter tolerances for wing shape, higher output engines run faster and are built lighter, and so on .
If you visit in person one of these production lines you see endless bundles of wiring being carefully hand routed and handled, many steps where documentation has to be performed before the assembly can proceed, and many places where production stalls due to a missing part or the supplied parts are defective.
Anyways all this is automatible with AI driven robotics. And it so happens that robotics are very similar in construction to high end aerospace products, so the robots could potentially manufacture and assemble more of themselves.
This would make possible numbers that can only be imagined in theory. Winning parties in future wars could have arsenals of the best weapons the tech base allows, and 10 or 100 times the quantity of the rest of the world combined. Not 10 or 100 attack submarines, but 10s of thousands, sufficient to make the entire ocean impassible. Not 180 high end air superiority fighters, but 180,000 or 18 million—impossible numbers and sufficient to cover all the airspace on the planet, making flight impossible for anyone but the winners.
Onboard ai need not even be advanced or better than humans, merely sufficient that there is a chance of beating a human pilot or warship crew sometimes. Overwhelming numbers do the rest.
There’s one major element for AI use in warfare you skipped.
Every modern war has been a contest of industrial capacity. More and higher end weaponry has proven to give a decisive advantage. But the more complex modern weapons are extraordinarily expensive to build.
Part of the reason for the high cost is a modern missile, torpedo, submarine, jet fighter, and so on uses far more components than past weapons, with many complex embedded electronic systems and many new ways for the weapon to be unavailable from an internal failure. Materials are also much more complex and difficult to build, with higher precision required as well. For example, compare carbon fiber to cloth or riveted aluminum sheets for aircraft wings. Higher speeds for the aircraft required tighter tolerances for wing shape, higher output engines run faster and are built lighter, and so on .
If you visit in person one of these production lines you see endless bundles of wiring being carefully hand routed and handled, many steps where documentation has to be performed before the assembly can proceed, and many places where production stalls due to a missing part or the supplied parts are defective.
Anyways all this is automatible with AI driven robotics. And it so happens that robotics are very similar in construction to high end aerospace products, so the robots could potentially manufacture and assemble more of themselves.
This would make possible numbers that can only be imagined in theory. Winning parties in future wars could have arsenals of the best weapons the tech base allows, and 10 or 100 times the quantity of the rest of the world combined. Not 10 or 100 attack submarines, but 10s of thousands, sufficient to make the entire ocean impassible. Not 180 high end air superiority fighters, but 180,000 or 18 million—impossible numbers and sufficient to cover all the airspace on the planet, making flight impossible for anyone but the winners.
Onboard ai need not even be advanced or better than humans, merely sufficient that there is a chance of beating a human pilot or warship crew sometimes. Overwhelming numbers do the rest.