> Just because you can find a handful of mistakes, or a thousand mistakes, or a hundred thousand mistakes, out of hundreds of billions of moderation decisions i

> Just because you can find a handful of mistakes, or a thousand mistakes, or a hundred thousand mistakes, out of hundreds of billions of moderation decisions is not evidence of evil or malice. It's not a handful. Also, who cares if it's evil, malice, or incompetence, it's happening and it is YouTube that is making the mistakes. I think you might also be missing the point, people are annoyed at that there's no way to rectify these mistakes or speak to a reasonable agent of the company more so than that they happen in the first place. Algorithmic moderation could be okay, so long as you have the necessary human operated escape hatches, since you are in fact impacting humans with your algorithms in very real ways. There needs to be some point that you can talk to a human with the power to be reasonable.