I was blackmailed – any YouTuber could be next

Instead of actually trying to create the best customer service, to me it feels that they treat it like a fun ML problem that they want to solve with magical algorithmic pixie dust.

Yeah I know it's terribly expensive to keep a horde of content reviewers, but then it would be nice for Youtube to be upfront about it that there is maybe a human in the process, but most likely not. I guess honesty is not anymore a valuable commodity in this world, yet it would feel nice sometimes to hear the truth, not some lawyer jargon to avoid all responsibility.

> "I also see a lot of people complaining that YouTube/Twitter/etc aren't doing enough to take down false/immoral/illegal content quick enough"

I dont want these companies touching false/immoral take downs. who are they to decide? sorry, I dont want these people in charge of the moral police. illegal stuff, like revenge porn makes sense.

> "YouTube is doing it's best to blend automation (fast but inaccurate) with human curation (more thoughtful but slower), and sometimes it gets it wrong"

I feel like they usually get it wrong, they are closer to being a blind man swinging a chain saw than a surgeon with a knife. HIRE SOME PEOPLE. they have huge margins, it would take very little for these companies to hire some people to review content.

> " feel like most of the time I see posts like this, the situation is resolved favorably and relatively quickly."

Tell that to people who have been shadow banned or demonized. most of the time, there never is resolution, their channel and all the hard work they have put together is ruined forever.

I understand there simply not being enough manpower on Earth to manually review so many cases, but the balance seems to be so heavily skewed towards automation (which still does a very piss poor job at it) that just feels like there's no human element vetting the final decision at all, and that's the approach which so many people have gripes about, rightfully so until Google strikes that balance.

I agree with you that it's a really hard problem, but I don't see this as finding a middle ground. In reality, there are problems on both sides (people posting bad content, and bad actors taking on good content). What I expect from Google is a constantly improving process on making sure bad content is removed, and good content is not. It doesn't seem like Google's process is getting better - if anything it's just swinging back toward blocking content faster. I feel like there have been a few swings, such as back in 2011 when any background music or movie clip might mute or block your video, even when it's fair use, to a more open model a few years later, and now we're swinging back. What I would hope is that they improve over time. And yes, many times I think that involves actual people looking at things, which is expensive in terms of people as well as having to deal with tough content (mental health of those moderators being really tricky, just look at FB). But the amount of arbitrary and poorly worded emails about how you violated the TOS but won't say how. It's obvious that working with people is more expensive than not, and that seems to be what they are doing.

It's always a cat and mouse game, I guess. And Google has fought this type of thing before with trying to fight people gaming search ranking.

Now people arguing over what content should be allowed is obviously a more tricky, moral, problem. But between illegal content and people doing illegal acts like blackmail to valid creators seems pretty similar.

They can defend against DDoS attacks, why can't they do better against fraudulent takedown requests? I think the answer lies in their monopoly status and the fact that creators don't have many other options for getting real traffic outside of YouTube.

If you're trying to make money from digital video, getting banned or even demonetized from YouTube is a death sentence on your vlogger career.

So sure, I don't blame YouTube. They're likely good folks doing their best to solve hard problems. At the same time, that this is such an impossible problem to deal with is probably a good reason to make sure Google doesn't have too much monopoly power over advertising (in this case video ads).

In short, the stakes are too high! They'd be lessened with more alternatives.

I think the best solution would be a wide, distributed network of automated takedowns targeting every nth video, but avoiding the high profile accounts. Ramp this up to a large enough issue and they will be forced to respond with a better process, otherwise their content network dies by their own process.

I'm not actually advocating that the above should happen, just suggesting it would turn the tables. Fight bots with bots.

Here's a thought experiment:

Imagine you are a global platform which needs to deal with inappropriate behavior (posting of copyrighted materials, etc.)

Flagging content requires a verified account, where you must authenticate your meatspace identity: a small monthly payment via a traceable payment mechanism; a text to your phone number; a postal address verified by sending you a postcard; a photo of your government-issued ID; and webcam photo of yourself.

None of this information is for public exposure! It merely ties your account to you, as a person, as tightly as possible. Nor is it required to use the service -- only to engage with the moderation system.

All of these can be automated from the side of the service provider.

For companies:

If you are working on behalf of a company, your company membership requires similar verification. You are also limited to flagging only IP-related issues raised by your organization (copyright infringement, etc.)

For individuals:

You must be endorsed by three other verified users.

False flags beyond a certain threshold will erode the weight of your flags, up to the point where your the moderation system will outright ignore you, with a periodic (automated) review to see whether or not your flagging behavior has improved.

Endorsing a certain number of false-flaggers will also cause your account to silently lose the ability to flag.

This mimics real-world behavioral modification for groups of people.

Thoughts? Where does this go wrong?

This is because the only time a post like this rises to the top on HN is because it's affecting someone notable in the tech/startup circle or someone with a lot of tech followers. If you don't have this kind of leverage to get Google's attention you're usually SOL. An example: https://www.youtube.com/watch?v=LrMvyVEQO6M

Edit: follow up to the above video: https://www.youtube.com/watch?v=zIK0yafnbGM

From the transcript:

> long story short it only went back up I got a message this morning from someone who is a fan of mine who is a Google employee who kind of reached out internally about it after finding out what happened and that basically immediately solved it after three months of me trying everything that I you know basically everything that I knew to try

So yeah. Unless you have a fan at Google, you can go pound sand if you run into a problem.

Automating thins is fine, and even getting things wrong is okay. YouTube is nothing without the content that's on it, so it goes without saying that any money-making channel deserves expedited support, especially in cases like this one.

If these organizations don't want to be responsible for the content on their platforms, then they shouldn't censor content unless ordered to by a judge.

They want it both ways - full editorial control, but no responsibility.

I've never seen a revenge porn video on Youtube, where are you coming up with this?

Just look at the robotic tone of their emails, they don't care. I've had a channel taken down and all my old videos deleted due to some nonsense. There is no appeals process.

This isn’t “cut them some slack” xor “revenge porn”.

This is the automated (mis)application of a rule against educational videos, a rule that should not exist in the first place.

Censorship platforms are inherently dangerous.

Isn't that exactly the complaint? There's no sane human review until an issue gets embarrassingly publicized--so if the victim can generate that publicity then the system "works", and if they can't then you never see the post.

Two aspects of the problems are very easy: first, the automated parts of the review system can be marked as such. Automated emails can trivially be signed "You Tube Bot" or something (such signalling should perhaps be mandated by law). Second, when a video has been reviewed and found "not guilty" once, it can be marked as such, and past reviews recorded. Preventing automated takedowns for the same reason, or requiring stronger reviews, is trivial (though more costly for the videos that do end up being re-reviewed).

They may have other reasons not to pluck those two low hanging fruits, but difficulty is probably not one of them.

That being said, there arguably are some contents that should be taken down. But content takedowns should always be done with care, and be used as a last resort. It certainly should not be taking down harmless gaming videos.

It's also disingenuous to bring up revenge porn to strike down criticism. No one could possibly mistake the video in question for revenge porn. Even their automated system didn't. It has nothing to do with this issue at all.

If it doesnt work on large channels then there's no blackmailing possible. If it works on smaller channels then you can't be a "bad actor" as easily.

At large enough scale (e.g. Youtube scale) there will always be people complaining about almost everything.

You can't please everybody, but it's in Youtube's long-term self interest to please the people that produce the content that keeps people coming back to Youtube, otherwise the creators will look towards other platforms (and we're already starting to see this).

> It really sucks when a legitimate video is taken down by mistake. But it also really sucks when revenge porn is left up....YouTube is dealing with two opposing issues

Except here, a third factor is in play - the extortion which caused a legitimate video to be taken down. Youtube's going to need to fix that problem fast.

And that's the exact reason why Youtube should be considered a platform, that Youtube is so viciously fighting now and all the Youtube believers are blindly accepting.

Not that I believe PragerU's videos are factful or informative, but I think the argument by the court that Youtube is not a state actor is completely missing the point, and it only shows how outdated the legal system is in the US, but it doesn't actually reflect on the actual situation, i.e. how much power Youtube is wielding, and how it is censoring and actively blocking free speech.

These kind of decisions shouldn't be granted to any private firms with such user and content size without public, democratic and transparent influence on the decision making.

BTW, all your arguments seems anecdotal to me. Sentences like "I see", "I feel", etc. don't really convince me and should not anyone else either.