The big lie of online censorship is that controversial cases where most people think the person "deserves" to be banned are unrelated and totally separate from cases like this one, where it's obvious the ban is preposterous to anyone possessing common sense. They are directly related. They are created by the same systems built under the same assumptions with the same mentality.
This is not going to be fixed by "better" algorithms, because it's not an issue with the quality of the algorithms in the firs place. The algorithms seem low-quality to you because you're judging them by a standard the company running them didn't use.