At this point, I'm no longer comfortable giving them the benefit of doubt. They burned that bridge. If this is a problem with aggressive filtering, they have had ample opportunity to note the problem and provide a solution that applies to all.

I really can't trust that they are benevolent or incompetent.

Does it matter all that much if they are incompetent or if they are maliciously censoring?

I'm actually not sure what is worse.

Indiscriminately punishing a small number of people for legitimate speech is at least as chilling as going after specific viewpoints -- any speech can get you banned, apparently.

An organization has developed a semi-automatic process that is consistently flagging documented evidence of various nations committing crimes. Human employees of that organization, when looking at the flagged documented evidence, are consistently deleting it. Assuming these humans are competently doing their job, they're following processes that were authored by the same entities responsible for (which is not the same as the implementers of) the machine processes.

edit: /implemented by/the implementers of

That would be fine if they didn't insist that the automation simply flagged it so humans could review it for the final decision.

Eh, human review isn't all that and a bag of chips. Firstly you have the subjective problems in human review, that different humans will find different things egregious. Secondly you still have the context problems: would the humans have been more likely to review the videos in the context of a reporter reporting on evidence presented in a court trial?

Thirdly, and perhaps most importantly from the standpoint of anyone building a system to moderate content, you have the human toll: most of the content the humans are reviewing is going to be highly objectionable to most humans, by definition. Expecting humans to review this content day in, day out and not have it take a psychological toll is not realistic; you can find interviews with people who have done this work for various companies, and the tales of their burnout.

(This is not meant to defend Google or condone what the reporter is going through, etc, just to point out that "make humans review everything" is not a slam-dunk answer.)

That appears to be what happened in this case.

> I appealed both decisions. Those appeals were apparently rejected, (at least that is what I gathered.because the videos were removed). That the community strikes stood (despite my appeals) was not explicitly clear during the process.

Is the appeal process automated? If there is no way to circumvent an automated process then it has to work 100% correctly, 99% is simply not acceptable.

Almost all people will appeal, it's quite easy. This brings us to the volume of appeals. Google has never liked to support this type of grunt work.