- a list of clear rules
- takedowns must provide a reference of which video caused the takedown, roughly where in the video it was, and which rule it was violating
Someone may reply that this is feeding information on the rule boundaries to spammers, at which point I make a second suggestion which could apply to more services:
- a "this platform is valuable to me" deposit, of say $10, which guarantees human review of administrative decisions.
It's people's livelihood. You can't just vaporise it. Well, you can, but eventually the angry mob you create will get large enough and come for your statue.
It would be difficult to structure this but simply allowing a video creator (or any fan of the channel!) to pay a Youtube analyst or a third-party auditor at cost ($10 seems reasonable for 15 minutes of even a decently-payed, English-speaking Western employee's time) to review the channel/video for 15 minutes for actual violations would be a huge boon to real content creators.
This would also allow Google to identify what takedown requests are blackmail botnets and block them from doing further harm.
If anyone at Google is reading this please consider it! So many viewers would be willing to pay that this need not even impact creators.
(Obviously this would have to be kept from getting out of hand and turning into Google racketeering with multiple takedowns of the same channel, but I trust them not to do that)
I don't really see how this is any different now. Google is already generating bad PR by not having an appeal process for these issues and plenty of people are already complaining that it's all a scam.
If a nominal fee gets you a written report that details exactly how you offended YT's policies and creates a possible scenario which you're given the opportunity to course-correct perhaps, then it's a win. What you're arguing would happen is already happening.
Right now, blocking a video will lead to Google being unable to earn money from it. Their incentives are aligned with those of the video creator.
> What you're arguing would happen is already happening
No, right now you are not pressured into giving Google money to review their automated decisions. With the suggested system implemented, you might fork over $10 individually for each of your 452 videos and get humans picking mostly the same useless text block in response, not really helping you to understand their decision. You will feel kind of like now, except you notice that YT actually made more money by banning your videos and start to wonder if that is by design...
Courts are great, but have to be independent. Even arbitration can be skewed in favor of bigger companies, since getting their recurring business is in the arbitrators interest.
I'd still be in favor of setting up something like that for lower level decisions, but understand this can only happen if Google finds a way to stay out of the line of fire. And still keeps a way for them to overrule those third party decisions...
I would say that's only true if the volume of video content was a bottleneck for their advertisers. They have more than enough content.
The given channel was estimated to make $4 per video. Just holding a small amount of revenue for review would pay for the human time.
Losing $10 is vastly better for the monetized creator than losing the channel completely.
When you issue a DMCA notice, you're certifying that the content is yours, under (theoretical) penalty of perjury. But big content studios erroneously issue them using automated systems all the time, and the only way to actually enforce this is for you, the individual getting targeted by one, to personally take legal action against a multinational corporation.
Meaning that, for all practical purposes, the people this affects have zero recourse.
Nearly 100% of people who receive a DMCA notice of alleged infringement (read: 'copyright strike') from YouTube have the option of filing a counter-notification.
In nearly all cases, YouTube responds to valid counter-notifications by informing the party that initiated the takedown. Counter-notification starts the clock on a 10-day window, within which the party alleging infringement must - if they want to keep the video down - sue you in federal court and prove to YouTube that they they have filed suit against you. Otherwise the video goes back up.
That procedure is almost always available. The downside, of course, is that you might get sued for an enormous amount of money that you may be held liable for. Statutory damages are scary. (But thy are a consequence of infringement, not notice-and-takedown.)
The only way to modify that behavior is to make it hurt.
Or the far more radical point of view that people who are economically dependent on Google are employees .. although that is probably impossible to define and unworkable.
On the other side, the content producer's "defending" karma should also increase each time the claim is rejected.
And then, the "content police" should somehow need more "work" to ban or remove something the weaker the claim is or the defendant is "stronger."
I am aware that the "platforms", however, simply don't care much now, simply prioritizing the "reaction."
For example, it is very hard to come up with a sort list of clear content rules. There are all sorts of ambiguities. For example, when is nudity ok? What about violence? Every line you draw in the sand will have false positives and false negatives. See any number of stories about Facebook moderators who despite having lots of content rules, will make subjective and contradictory judgement calls.
Furthermore, with a concrete list of rules, bad actors will act just to the left of rules violations. Clearly, having some feedback is required. In this case, however, we don't know what Jordan Pier knew, since this was initially a reaction to the channel's deletion without consulting with Jordan about what happened.
Finally, having a support deposit--no matter how small--will disadvantage many creators and increase the divide between haves and have-nots. I admit that it's frustrating that you literally can't purchase a human's attention, but every policy decision like this has profound ramifications when you have billions of users.
These aren't new questions, organizations like the BBFC and MPAA have been wrestling with them for years. And Google will have been wrestling with them internally .. in total secrecy. Yes, there are always going to be edge cases and people intentionally working up to the edge.
> having a support deposit--no matter how small--will disadvantage many creators and increase the divide between haves and have-nots.
I get that, but at the moment everyone's disadvantaged, and the larger creators potentially for tens of thousands of dollars.
Secondly, they work with (largely) mainstream content producers (e.g. movie studios), who rarely act with downright malicious intent because of reputation and money on the line.
I don't believe content moderation is a problem solved well anywhere.
Joke's on them: Google Maps will just give the angry mob directions to the Amazon HQ instead.
If this was a hardware or software issue it would be obvious that you're trying to scale systems that do not scale.
In technical terms, social "algorithms" are O(k^n), exponential.
You're trying to do the impossible, eh?
(Really, the argument could be made that FANG et. al. are becoming a new de facto world government. From that POV you're attempting to solve the hardest human problem, eh?)
- - - -
No one is forcing you to be that big. At the end of the day, scraping away all the bright-eyed rhetoric about "connecting all the people, places, and things for great good", you're doing it for money. Building a moderated social network isn't impossible (maybe) but it's woefully expensive.
The existing systems are, arguably, harming people and messing up society, tech companies notoriously have a lot of money already, but are unwilling to prioritize and pay for the moderation that could (arguably) improve things.
So... what? I don't have any answers, I just want to point out that from my POV the problem is clearly hubris and basic cheapness.
FWIW, I don't think money is the limiting factor here. If there was a known good solution, but it took $10B to implement, I think it would be done.
You could pay X thousands (more) moderators, but there are drawbacks in a human solution as well. First is the fact since there's human judgement, rules will be applied inconsistently, and probably with some bias.
Second is the fact that moderation is a very mentally exhausting effort. You can observe some of the cyclical nature of this balance from the news over the past few years. Article about moderators with PTSD? Well, automation is the solution. Automation fails in a lot of edge cases? Well, why not hire armies of people to do it for you?
I don't think there is an obvious solution here, no matter how much it cost.
The issue is centralizing social decisions to a single entity. We already have a model where corporate entities provide services at scale without taking on much of the social aspect: Utilities and infrastructure. So find a way to become a video hosting utility. An uncaring pipe. Create tools for users to build their own fiefdoms and branding and have it moderate themselves. Tying advertising into that might be difficult, but perhaps it would be up to those users, pay for the services or automatically enter an agreement with advertisers. If advertisers pull their content for whatever reason then people could choose to pay themselves or perhaps make a deal with different advertisers.
Hosting adult-related content that most brands don't want to touch? Perhaps make a deal with seedier advertisers. Youtube would still take its cut, but only as part of infrastructure fees.
Want to promote your content to other circles? Well, you're essentially an advertiser yourself now.
If that's the case, then it may be back to Youtube doesn't scale.
> Second is the fact that moderation is a very mentally exhausting effort.
You have an impossible trilemma. Youtube:
* Allows/must allow anonymous or pseudonymous content creation, without the ability to ban bad actors;
* Faces significant social (sometimes legal) liability for content hosting and especially curation/promotion; and
* Wants to provide friction-free, speedy, and consistent policies to its content creators.
Those can't all exist at the same time. Right now, what gives is a bit of #3 (seemingly arbitrary, severe, and algorithmic actions against channels) and a bit of #2 (occasional stories about how Youtube is evil for promoting flat-earth Nazism vaccine-denial theory to children or whatever).
Youtube could go with a heavy-moderation approach, by requiring content creators first register with a government ID (plus business license in each country of intended audience) along with prior human-moderator approval of each uploaded video. But that would fundamentally change the platform into something more like a traditional TV network rather than dynamic social media.
Other commenters here go heavier on the "free speech" side of things and think that Youtube ought to face no repercussions (social or legal) for hosted content, freeing it of the need to moderate. I'm not sure how we would get there from here, but nevermind that job for lobbyists.
Society as a whole seems to implicitly solve this problem by not being a big giant curator under its own brand. Traditional media still exists as many (notionally-)separate outlets, each with their own branding and editorial views to offer at least limited competition. But if this is the only viable model, it's still another way of saying that Youtube doesn't scale.
The critics think that they know better than society and impose demands not codified in any law nor how things remotely work. How the hell should YouTube have decided to keep its size acceptably small then? Youtubes 1 to 99999 and hope people go to the right arbitrary subsegment? The demanda are arrogant and clueless as saying that builders should only use rubber headed hammers and nailguns should be banned because it would reduce construction noise knowing nothing about the field or its practicalities. Now /that/ is hubris.
> If they were 50% smaller, which 50% of their current content do you think they should get rid of?
Couldn't they just split in twain? https://en.wikipedia.org/wiki/Cell_division
(I lied when I said, "I don't have any answers". I do have answers.)
- Add a "reputability score" for every channel which is log((total videos) * total views * total likes * months since first video). Obviously there should be a HUGE burden of proof when deleting a channel with millions of views and over a dozen videos.
- These are just two 15-minute-ideas of an engineer who can't get an interview at google. I'm sure among the thousands of employees somebody has an even better one.
YT has policies against animal cruelty, but a search for "electric mouse traps" shows a bunch of live animals being killed. (There are a lot of these).
https://www.youtube.com/watch?v=w0D4rJ6zNuM
https://www.youtube.com/watch?v=oh_QXnJcP7M
https://www.youtube.com/watch?v=qwKN8h1du8s
https://www.youtube.com/watch?v=Dgh9Bs4TFdY
Tar paper traps lead to a slow death by starvation but it is a side effect of the function - to work without "triggering" and inconveniencing anything much larger instead of risking harm. No need to take a cat to the vet if they step on it, just wash them off and maybe clip some fur keeping it attached. The fact it causes more suffering than a mechanical trap isn't enough to qualify although one may object to its usage. But breaking a mouse's leg just to laugh at it and watch it crawl? Gratitutous and would qualify as animal cruelty.
And you should blame that (that nobody flagged them) on the double-standard most people have towards "not hurting innocent animals" vs "freedom to do any slaughter with the so-called 'pests'".
As I say I don't mind the killing of a healthy animal for meat, but the milk industry just sticks in my throat (especially when it is so easy (and tasty) to cut out milk these days)
(or was your comment a joke :p)
This discrepancy between how we pretend to protect animals but still do plain "pest control" is weird (this is a weak word), but this is a quirk of our "modern" society.
The law is made democratically, so that just means that a majority of the citizens think that it is not ok to hurt animals, but that it is ok to use mousetraps. Blame on the lack of awareness of the society.
But then, in turn, this means that not a lot of people flag the video, if the majority of them still think that pest control should be legal. If not, then now should be the right time to change the law of the country where you/me live.
You can't just ban everything with electric mouse trap in the title - this video [0] is on the first page of searches and doesn't contain any rule breaking content. If a band appears titled "electric mouse trap", do you want that content to be removed?
There are so many topics, so many videos,so many people actively trying to skirt the rules that it's just infeasible to moderate purely by scale. According to [1], there are 30k hours of content uploaded to YT every hour. How do you suggest they moderate even just the new content, never mind the backlog of content.
To be clear, I'm not defending the content on YouTube, or their moderation/takedown practices and responses, I just want to point out that it's not as simple as "just moderate better"
[0] https://m.youtube.com/watch?v=GYIk14jyXzw
[1]https://www.statista.com/statistics/259477/hours-of-video-up...
> There are so many topics, so many videos,so many people actively trying to skirt the rules that it's just infeasible to moderate purely by scale. According to [1], there are 30k hours of content uploaded to YT every hour. How do you suggest they moderate even just the new content, never mind the backlog of content.
That's really an argument that they have to reduce their scale until it matches their moderation ability (or increase their moderation ability until it matches their scale). We wouldn't accept a car company failing to recall and fix defective cars that they sold because their repair shops lack the capacity to do so.
Send 999 out of 1000 viewers away? Delete 999 of 1000 channels? Block 999 of 1000 video uploads? I genuinely don't understand what you are looking for.
And note that this would still be 30 hours of new content per hour. And it would change nothing about the economics of moderating content.
The only content policy that matters is unwritten: "No content which could cause a large or loud enough objection to disrupt add revenue."
Objections to mouse traps or hunting is not large or loud enough to hurt the bottom line.
If you mean that the channel didn't do anything "wrong", and shouldn't have been deleted, I totally agree.
Moderation here isn't the problem. Youtube's policing has downright enabled censorship on the site.
I don't think anybody here is forgetting the scale; they simply understand that Google-scale problems come with google-scale revenue.
Omitting this fact seems like an insincere overture to the community of upset users, to me.
The channels that do actually have revenue impacts are the ones with actual people reps already.
https://www.youtube.com/channel/UC40wiRfP8u2DrVvUm_VGcqw
https://www.youtube.com/user/CraigTube/videos
Archived copies:
https://web.archive.org/web/20200227065130/https://www.youtu...
https://web.archive.org/web/20120625045947/https://www.youtu...
Saying what video and when in the video tripped what part of the magic algorithm is also doable. I think we both understand that the algo has to do the work, and it has to output a "yes/no" at the end. However, absolutely nothing forbids coming up with some kind of feedback or error message. The algorithm is doubtlessly complex, and was therefore debugged. Yes/no was not sufficient for debugging in its development. The mechanisms are present.
Expose this to the user.
YouTube is an internet advertising company like pretty much everything else Google. That they serve video content is ancillary.
Maybe perform fewer moderation actions?
There are many sophisticated actors that try and game the system, and it's because of them at sometimes the system makes the wrong call.
And before you say this won't scale, this is how Reddit moderation works, and seems to work for millions of posts made every day.
Also, some sort of public log of takedowns and reasons in some form we could verify might also be nice. It's super creepy seeing someone "disappeared" from the Internet without so much as a peep from Youtube/Google.
People should not be expected to somehow get another famous youtuber and a hang of an insider through HN front page to get their problems resolved.
Well this is BS, it's like saying the author of a book at a library should be the one to complain if the library censors the book. Don't the library's patrons get any say?
If Google can't operate at scale, then maybe it just needs to be shut down so maybe someone who can will take their place.
Google simply doesn't WANT to operate at scale because that will impact profits.
Then don't take it down, and let the Justice Department do it. Or accept being an editor. You can't have it both ways.
Your appeal to scale here doesn't apply to this case.
It's possible (however unlikely) that there was a justified reason for the take down. Has anybody heard from the original creator?
Is there a reason why Google cannot hire more people?
I consider adding that to every Google related comment, but it gets tiring.
It does certainly seem like this was an erroneous take-down. I'm not familiar with the content, but I can't imagine a more innocuous channel. If JPdylon is out there, the first recommendation is to appeal the decision [1].
Granted that doesn't always work, and it's unfortunate that common next step is to resort to "Support via Hacker News". That being said, content moderation is a super hard problem. It needs to get better and more fair on both sides (i.e. taking down what should be; leaving up what is quality content), but the scale here is often forgotten.
There is far, far too much content posted to manually review. If EEVblog's suggestion was taken and Susan (YouTube CEO) had to manually review every moderation action on channels with 5K+ subs, she would not be able to keep up. If you had a team of people, they would not be able to keep up. Even if you had a fleet of people, the scale is just unreal.
EEVblog's suggestion to diversify platforms is reasonable. The fact that those platforms don't have some of the same problems doesn't necessarily mean that they're better at content moderation, however. They just don't have to operate on the same scale.
[1] https://support.google.com/youtube/answer/185111