To Combat Drone or Not to Combat Drone?

[Ritterin Sophia]

Didn't notice this earlier.

What bothers me is that the concern is being applied to the near future drone generation, by a public that thinks "a robot is a robot" and cannot tell the difference between next-generation UCAVs and Skynet.

That creates a fear of automation independent of the evidence: a fear that could be rational in principle, but only if the systems being proposed were different, and were smart enough that there really was a threat of them actually going rogue, rather than merely screwing up on the tactical level due to a software bug.

Tactical screwups are bad enough, of course, but they're a whole different level of problem than the danger of a general AI rebelling.

For Valen's sake, could you you try not strawmanning the position!? It's not that we're worried about skynet, it's that we realize the infeasibility of making UCAV with sufficient target discrimination capabilities to determine whether a formation of people is a refugee group or a company of infantry whilst flying high and fast enough to not be vulnerable to AA Guns let alone missiles after it gets disconnected from it's control satellite. As Mark has shown it's easy to take a satellite out of commission or severely impair its performance, once you have orbital lift. With technology proliferation as it is this is only going to become easier and easier for less wealthy states to do.

[Starglider]

It's not that we're worried about skynet, it's that we realize the infeasibility of making UCAV with sufficient target discrimination capabilities to determine whether a formation of people is a refugee group or a company of infantry whilst flying high and fast enough to not be vulnerable to AA Guns let alone missiles after it gets disconnected from it's control satellite.

Never mind UCAVs, it's impossible for an F-16 pilot to make that distinction from 20,000 feet with no intel from HQ. If you want to drop stand-off weapons on infantry, in the absence of strategic surveillence resources, designation by friendly troops or at the very least a competent forward air controller is essential. The challenge is getting UCAVs to respond to local commands (of the nature 'hit laser-designated target in grid 3R with 500lb bomb) without creating a security problem - but once that is achieved they will probably do this more reliably than a human pilot would.

In any case when are you going to be doing deep strike against enemy air superiority and jamming, and then trying to target random columns of infantry? That part comes in the mopping up phase after the enemy's air force and hard-target-based defence infrastructure has been obliterated.

[Simon_Jester]

For Valen's sake, could you you try not strawmanning the position!? It's not that we're worried about skynet...

Apologies. You're not.

Some other idiot is. That is what bothers me. It's not the realistic criticisms (can this hardware do the job reliably enough?), but the unrealistic ones (will this hardware lead to the Robot Uprising?).

The concern about reliability is realistic, and therefore reasonable, and I have no intention of trying to attack people who are worried about UCAVs for the kind of reason you are.

In any case when are you going to be doing deep strike against enemy air superiority and jamming, and then trying to target random columns of infantry? That part comes in the mopping up phase after the enemy's air force and hard-target-based defence infrastructure has been obliterated.

I agree with this, and it ties into something I said earlier: if the enemy can still interfere with your ability to talk to your drones, you haven't won the war yet, and it isn't time to start going after targets of opportunity yet. You can plink tanks after you take down the command posts, the airbases, the radar installations... the stuff that can interfere with your ability to do tactical bombing with manned or unmanned aircraft.