I've never seen you play anything but Brawler, although my knowledge here is limited and you might play other mechs a fair bit as well. If not, "foregoing using AC" sounds humorous.
I've been maining Brawler since the end of TPG 2016, and even now I'll rotate to Assault, Reaper, Sharpshooter, Berserker, and occasionally Grenadier when the situation calls for it. When Tang picked me up in 2016 I was maining the Assault, Grenadier, and Berserker.
Someone who ground-dodges intelligently achieves exactly the same thing. You can angle-dodge forward and backward, and you can jump-dodge with structures (e.g. jump on a bridge banner on Prosk and immediately dodge from it, doing something quite similar to an air-dodge). So, I'm not sure how this phrase contributes to your argument. Someone angle-dodging on the ground while fighting an opponent on the same elevation is still going to be traced primarily in one dimension (x-axis). Assuming the ground is flat, the angle-dodge still doesn't present the same variability in being traced as jumping in the air and dodging to the side does, especially if the enemy is close. If someone uses an air dodge, then additional dimensions of tracing are introduced into the engagement.
I agree with your argument that others can use the environment to achieve some elevation change, but those environmental factors aren't present everywhere in a map. For example, a number of players equip AC on Origin simply because the hang time from using jump pads can be dangerous.
Barely matters if you aim in between 1 and 2 mech heights, which is what I do if I don't know what to expect from my opponent, which is not often the case. I think it absolutely matters. When I'm about to get into a corner fight with a good player, and I know they run AC, then I know I can't just set my sights right next to the corner and wait for them to poke their head out. They may boost jump around the corner, so now I have to aim higher if I want to connect with my hitscan weapons. I can still launch a remote detonating projectile like a grenade or TOW, but the splash damage is going to be less than a direct hit. Sometimes a specific radar pattern will give them away, but if the timing between the peeks is too tight then looking at a radar might cost you if you miss a shot.
I've faced very good SS, Scout, Assault, and Infil players use AC in this very exact method to good use. Sure, that tactic may not work every time, but it works well enough that it's a tactic I see regularly with mid to high tier players.
People tend to overuse this unoriginal move and fall victim to its illusory effectiveness in high-mmr lobbies. However, it works wonders below 2200, which teaches people that using it all the time is beneficial, so they do just that. I think it's a bad idea to remove an internal based on the argument that it doesn't prepare them well for high-tier play. Which, by definition, means that the majority of the population will not get to that point, so I don't see why an internal needs to be designed in such a fashion that it nudges new players over to the "correct" meta.
Every learning experience in every game is based upon identifying, often by trial and error, what "works" and what "doesn't". And for many players, what works in the early stages of a game against inexperienced players will not work against more experienced players. It's like how everyone got upset with Bastion in Overwatch with how quickly it slaughtered the noobs, but eventually the population learned the limitations of Bastion and overcame it.
As for all the talk of bad habits and unoriginal moves, that's irrelevant to me. I'm not interested in categorizing the lameness or originality of a keyboard combination. Plenty of people have put the time into breaking the bad habits of using AC to hang in the air for eternity, and what you call bad habits I just call "lessons in progress of being learned". Let them screw it up. That's how pretty much all of us learned how to do anything better at this game, including when to use AC or not.
Overusing it in lower-tier matches is almost always effective, which, again, teaches people the wrong thing. I've seen this first-hand countless times. Again, I don't think it's a "wrong thing" to be taught. People who use AC a lot may eventually learn when to use it and when to not. Some do, some don't. I see it being no different than failing to learn how to stay behind cover, or follow your teammates, or stay on the AA, or don't repair in a firing lane. That's like saying "remove the barrier or shield because it teaches people to take cover WHERE THERE IS NONE" I don't see why we need to take away the opportunity for people to self-improve and separate themselves from the herd.
Removing an item so that newbies don't learn a bad lesson is a slippery, slippery slope. Anyone here can use that logic to hilarious effect.
Might be true, but just because a mechanic increases the number of available options and utilities, thus increasing the skill cap (ostensibly), doesn't mean it is a good addition to the game. Applied to movement, not any internal that adds movement options is necessarily positive. Imagine an internal that lets you fly up indefinitely, until you run out of fuel. Or one that lets you back-dodge. Or one that changes shift+s to 90 degrees from 180. I agree in principle that not every additional mechanic is a positive addition to the game. On the other hand, we're talking about something a bit more complex - taking an internal that grants its user a movement mechanic it otherwise wouldn't have. I'm of the opinion that we shouldn't have internals that grant an additional degree of movement, and that air-dodging can be "safely" integrated with every mech so long as it was done in tandem with balancing air-compressor related values like fuel consumption. Having an air compressor internal is like having a boost or a dodge internal to me. If it's a movement mechanic, make it either universal or mech-specific (like Rocketeer, Bruiser, and Scout).
Put another way, I suppose that I'm just drawing a line in the sand with how Internals should be handled, and I don't think Internals should grant additional modes of movement. As I said earlier, I'm open to being convinced otherwise.