If a company chooses to automate something that should not be a defense. They should still be held equally accountable for their actions no matter if they employ a human or an algorithm to do their censorship for them. If they know their software/automation is shit and keeps screwing up, they're still making the choice to continue using it.