Nice toys, but old news...
'Networked UK battle sims are relatively cheap because they don't tilt and swivel like normal flight simulators' isn't exactly accurate - NONE of the fast jet sims have full motion systems because it just wouldn't do anything worthwhile - it doesn't give the right cueing and it couldn't give the right level of G input anyway.
A G-cueing seat (positive/negative/lateral G cue simulation), hookups for the pilot equipment (suit etc) and a dome visual is more than adequate to provide all the realism you could need in fast jet sim.
A motion system for a conventional full flight sim is necessary for something like an A320, or a cargo or surveillance plane because you don't have the suit, seat and dome visual to give you the cues instead.
And just for the record, missing out the motion system doesn't save any money & make the sims cheap, as the rest of the device more than makes up the difference. Indeed the fixed base fast jet sims are MUCH more expensive than the normal A320 types, because they're much more sophisticated.
As for the networked sims in this case, you'll find they're generally low fidelity devices compared to the 'Full Mission Simulators' used for the majority of training, and which can be used for all sorts of training and rehearsals. The low fidelity sims tend look like the kind of thing an enhusiastic DIY person could knock together with basic mocked up cockpits (and 13 projectors? Try 3), the high fidelity ones can cost the same as the aircraft and have full, working representative cockpits (and do have high projector counts).
Networked sims have been in use for years, though usually within the same site/country. The biggest problems with wider exercises are the sheer cost of inter-site links (the fattest dedicated link money can buy is needed to meet the data transfer and latency requirements), and the plain fact is that no-one wants the security headaches of linking their simulators and related databases together - this latter one is what kills most multinational sim networks, so they tend to only be done for test purposes with reduced functionality eg.“First WAVE”.
The simple fact with simulators is they always contain the full, real performance data for the aircraft itself, plus all the performance data that exists on the other equipment and aircraft that the aircraft will be used with/against. And this kind of data is highly protected, and not the kind of thing casually shared between nations - even with your friends you don't want to let them know everything about your kit, or let them find out just how much you know about theirs.
Anyway, it's all good toys, and much more entertaining to use than most games, though the cost is huge and obviously learning to work a 'properly' simulated aircraft isn't a 2 minute job unless you already know how to work the real ones. Most consumers would probably give up if they tried it as a game. And they'd run a mile when they got the bill!
But it does at least provide the opportunity to try out weapons and tactics in an environment where if it all goes wrong the only thing you need to do is restart the exercise, rather than clean up a smoking mess. Or deal with an enquiry about why you just killed that F22... ('sorry, wrong button!')
Plus the opportunity to do mission rehearsal is always useful - always better to find out the plan doesn't actually work before you're finding out the hard way.
As for the final point - autonomous aircraft are perfectly possible as even the current civil airliners can takeoff, fly a route and land automatically: the pilot is only there as a trust thing for the passengers, and to deal with equipment problems.
And the current generation fast jets are even easier - the pilot only provides an input to the computers flying the plane so can easily be replaced by a remote and/or programmed command system, the outcome of a crash is less of a problem, and there aren't any passengers to get worried. Indeed I believe that part of the future development path for the current generation includes updates to autonomous use, and this update is supported within the current avionics code.
Quite where this leaves the pilots, and all the training facilities is obviously a good question - somehow I suspect that while drones are useful there'll always be a need for a pilot in a cockpit, if only to deal with the situations that no-one thought of.