i dont think you will ever get rid of human pilots.
look at star wars Grin
That's actually the direct result of Burnside's Zeroth Law of Space Combat.
I'm still not convinced a pilot is 100% obsolete in the near future.
I also don't remember admitting that the AI would be making tactical decisions on its own.
I did realize, though, that I was mixing up my arguments between AI-craft in real life, and AI-craft in this game. In real life, having a human on station is much more appealing, but they'll probably be deferred to ground control or forward operators, leaving the risky stuff to the bots. You'll almost definitely need a human in the chain, though, for some jobs like being a forward air controller for close air support of troops on the ground.
But back to game-based AIs.
A FUO need not have an AWACS-sized craft. Modern AESA radars are probably enough to handle the data link so a one- or two-seat fighter is probably all that's needed. It doesn't need banks and banks of circutry, especially if most of the commands are simple vectors and instructions. (Probably a lot like an RTS) Even if the back-seater needed to see a live TV feed and directly control one of the UAVs, he can easily have the controls to do so with current technology. There's already talk of using the F-22's AESA radar as a datalink supplement, and using them as AWACS.
A fighter travelling at Mach 4 can cover 12 km, the range of the Sparrowhawk, in 9 seconds. Modern infra-red search and tracking systems, like the one on the Mig-35, can provide solutions for airborne targets at 15 km. Furthermore, at this stage of a merge, both the human and AI pilot would be looking for the target in a rather narrow area, knowing the target's altitude and bearing. In all the art of UFOs they've had nice, bright, white-green plumes from their antimatter engines, possibly increasing the range at which you can spot one. Any advancements in telescopic imaging will benefit humans with helmed-mounted sights just as much as an AI.
Assuming a head-to-head merge, at a range of 15km, and both craft at a speed of Mach 4, the craft will pass each other in 5.5 seconds. Let's assume that the missile has to be launched at least 2 seconds before the merge, so the pilot has 3.5 seconds. That's more than enough time for a trained pilot to get a missile shot off, especially with helmet-off-boresight technology - all he has to do is look at the target to lock on to it. In fact, it would still be a waiting game for the pilot: Well before the enemy gets into range, he'd have his HOBS activated, weapons armed, and thumb on the pickle, waiting for the tone that lets him know the weapon has a solid lock. After that, you're looking at around 140ms or less for the reaction time of a trained human responding to an audio signal. (
http://en.wikipedia.org/wiki/Reaction_time).
This is what I mean that the AI will not be significantly faster. In the 140 milliseconds it takes for the human to press the trigger, versus the AI launching the missile right away, the craft get 380 metres closer to eachother.
3) Ahh, I forgot about some of those. There's definitely a scope on the kind of decisions those can make, but definitely for things such as accidentally shooting airliners (image recognition) an AI would beat a human on the draw.
4) In the scope of the game, I realized one kind of luring into traps (SAMs) would be done on the geoscape. In dogfights, it depends on which combat model we're using. The current model is just lobbing guns and missiles at eachother from thousands of miles until at least one craft blows up. Future models will probably involve much more depth and strategy. Alien psychology has very little to do with things like that, either: Either the alien follows the bait or it doesn't. Using strategy is also even more important when you have the disadvantage in technology, too, because you can't afford to use more 'conventional' tactics.
How many pilots have to get killed before one gets an opportunity to exploit?
I could say: "How many millions of dollars worth of aircraft do we have to lose in engagements a human could have taken advantage of?"
Still, your argument will happen more often.
I expect the ejection seat is probably the heaviest part of human-specific fighter components. The second heaviest would be armouring. Life support and pressurization equipment is shared with the engine, and actually need not be too heavy.
Backup systems: Ahh, you're right in that getting rid of these will save a lot of weight and cost. The question then becomes how much are you spending on your AI fighter? A lot of those backup systems will save your craft from crashing from non combat-related problems, as well as combat damage. Better capabilities, like speed and maneuverability, also add to the cost, not to mention the sensors (probably the most expensive parts.)
The wasted space from a cockpit and canopy isn't all that much. In fact, if you tore off the canopy on modern supersonic fighters and filled the cockpit in so it were flush, but left the rest of the plane as it was, you'd actually
increase the drag of the fighter at transonic speeds, due to the area-rule. It's easy enough to design around that, but I thought it's kinda interesting.
It sounds like the kind of craft you're advocating is cheap and expendable, designed to only last a few missions. The sensors and computers would probably be mounted on pods that could be jettisoned for recovery, and the airframe would not be protected against fatigue. I don't know exactly how to make the engine, required for the high speeds, inexpensive enough, though, and the ability to sustain high-g turns can't come cheaply, either. I'd probably still put backup systems on my AI-only machines.
As long as we're talking specific to the game, though, I definitely can evoke Burnside's Zeroth Law. If players can track specific pilots, then they have the ability to develop attachments to veterans just like the ground troops. Specific human pilots can also level up with experience, unlike a learning AI whose upgrades will spread to every fighter nearly instantly.
I'll also point out again that none of the craft at the beginning of the game were designed with knowledge of the aliens' capabilities in mind, and the Stiletto is the only craft in the world designed with the possibility of going up against ETs, and it's 60 years old. There's also a proposal by BTAxis in the wiki about a better dogfighting system (which I love, but I'm going to post my own topic suggesting some variations), so we can expect the current system, based solely on attrition, to disappear.
