AI technology isn't something that actually exists today (outside of a few limited parlour tricks). It's science fiction. Procurement decisions have to be made based on reality.

Remote operation is vulnerable to jamming and anti-satellite weapons.

>You could have a plane that goes from point A to B and shoots down anyone that tries to stop it along the way without any remote operation.

And you don't envision any problems with this?

Does it matter what I envision? It's possible. It might be useful. So it will be done.

>shoots down anyone

Like Iranian or Korean Boings?

Sure. Just like humans tend to do. Or not if the right person cares enough to program it not to.

That is not what the link says it did not pilot even simulated jet. It issued orders as a controller would on border of AWACS I guess.

I don't think that's the case. Another link: https://www.popsci.com/ai-pilot-beats-air-combat-expert-in-d...

That's a constrained simulation, not the real world.

Can you indicate where in the article or the linked study those constraints are indicated? Military flights sims tend to be pretty decent for this.

The whole scenario was unrealistic. Simple rules of engagement, pre-programmed objectives, little or no EW, no ground control or AWACS, no ground defenses. The technology holds promise and will eventually be capable but we're still decades away from something that can replace human pilots across the full spectrum of missions.

The AI was defending, so it would've benefited from ground defenses, and the study indicates the human team had AWACS assistance.

Yeah, I think the big question is which style of C&C is better. A single pilot in a F-35 closer to the action, or a bunch of pilots further away in a sitting duck with a big antenna.

The answer might even change as AI develops.

If I had a budget the size of the US military, I'd want both.

I've never understood how the F-35 pilot is supposed to command both their own plane and an army of drones at the same time.

Once all the LEO satellite communication projects are up and running you could be pretty far and still get decent performance.

I think directed-energy ASAT weapons will be deployable around that time frame, so I wouldn't bet on the resiliency of LEO/MEO constellations there.

You're going to need a lot of ASAT firepower to put a dent in StarLink's constellation (thousands of minisats).

And thus lasers, which will presumably have a far higher sustained fire rate / magazine depth than ASAT-capable missiles. (And 100kW-class lasers aren't science fiction anymore, since the US Navy is deploying them for testing)

Jamming