Nothing of substance is "proven" by a computer. There are things that can be verified by exhaustive computer search. The proof part of the A&H theorem was the reduction to the cases that were then "verified" by computer. It was still viewed with suspicion.
< snicker>
So...it's a proof, and it hasn't been contradicted, and yet, it's "viewed with suspician"....sounds more like a blundering-around science than a pellucid palace of logic to me.
Well, I don't exactly know what "substance" consists of, but I helped write a generalized computer proof searcher back when I was an undergrad at Berkeley, and cranked out some marginally useful analytical geometry theorems that had not, until then, been discovered, as far as the math department, after substantial research, could tell.
At any rate, in what sense can you say nothing is proven by computer? My perspective is that the shoe is on the other foot--as long as you are merely bench checking proofs with hundreds of human eyes, you aren't proving squat--as the 300 year, off-again, on-again, off-again "proofs" of the 4 color theorem demonstrates. Only when a computer checks it over, do you have anything approaching a reliably vetted proof.