Earlier quoted context omitted.
The "but you can request a review!" handwaving all these massive tech platforms use to assuage concerns about heavy-handed (and coldly automated) policy application is really long overdue for serious consumer protection regulation.
I like to imagine the AI spits out a line like this: > Should we ban this user? YES Then you request a review, and the support person double-checks that the output really does say "YES" and not "NO". Then they tell you they reviewed their decision.
I wonder if someone has already tried to invoke it against a Google/FB/.. ban hammer.