It's not a minor change... Parents are to be responsible for their kids. I never disagreed with you on anything more than this.

This.

Oh...

I thought they took Youtube down.

Oh...

I thought they took Youtube down.

At this point, I wouldn't even care if YouTube shut down.

At this point, I wouldn't even care if YouTube shut down.


I would care.

And y’all thought this would be the worst thing that would happen this year. ;D

Has anyone noticed that it is now TWO ads screened before you get the chance to watch you required video?

It didn't end up happening I guess.

Has anyone noticed that it is now TWO ads screened before you get the chance to watch you required video?



No, I don't see any ads.

No, I don't see any ads.
How did you arrange that?

It didn't end up happening I guess.


I beg to differ:

SwFjj2Zg0NU

In short; most of COPPA still hasn't been administered. What few policies that has, resulted in a schism between 'established channels' and 'unestablished channels'. Channels with 1M+ subscribers saw sustained growth in the last two months, while Channels with -1M subscribers saw a decrease in growth.

How did you arrange that?


I don't know I guess it just happened.

It didn't change anything (important), it's still the same YouTube.

I don't know I guess it just happened.
May be down to the country you are in?

May be down to the country you are in?


They are pushing more ads because they want you to buy YouTube Red. I would just use an adblocker, I don't want to give Google any more money as it is.

They are pushing more ads because they want you to buy YouTube Red. I would just use an adblocker, I don't want to give Google any more money as it is.
Is the adblocker free?

Is the adblocker free?


Yeah, I use uBlock which is free (supported by donations).

Youtube has been crap since Google took it over!!!!!!

Agreed!!! I still follow some people on there but youtube has really gone downhill. The ad amount is horrible! I got premium cause it was that annoying. I give these f***** my money in order to watch the people I subscribe to without those damn ads.

Yeah, I use uBlock which is free (supported by donations).
Many thanks

As for the comments section.... that WOULD hurt creators but not just creators, followers as well. Interacting/chatting can really help people like me. It’s a way to meet people and have nice and interesting conversations. Just like this website. I truly enjoy talking with creators and other followers in the comments. Especially those live streams. It feels like I’m hanging out with them. As an introvert that’s a very beneficial thing for me.

May be down to the country you are in?


I live in the United States as you reside in The U.K.

Yeah, I use uBlock which is free (supported by donations).



I use Adblock Plus.

A Perfect Example:

As to why COPPA...... is sh!t

Ltw82AC1OZc


How in the world is something like this considered for kids? I used to watch this particular YTP 10 years ago, it's anything but child friendly.

Don't you have a degree in computer science? :P

Machine learning isn't 100% accurate. Report the video to improve the algorithm.

Don't you have a degree in computer science? :P

Machine learning isn't 100% accurate. Report the video to improve the algorithm.


Part of why I was never that keen on this policy in the first place ;D. As I said before, YouTube should've just lobbied Congress for newer regulations that would have made more applicable sense for the modern era we live in.

Despite that though; especially when Machine learning is still years away from being perfected, it's impossible to truly regulate what is and what is not exactly child friendly content. That was the main fear of many content creators, and this obviously inappropriate video being deemed 'for kids' sort of vindicates those fears. It's only a matter of time until YouTube execs begin to have a committee of actual overseers that determine content (like the one I posted above) on its likelihood of getting entangled in an algorithm that a child would likely fall into. But as we've seen with previous 'Adpocalypses', it's not entirely clear which route is rather preferable, especially if you're like me (and others) whom are not that keen on the draconian actions the platform has taken in recent years, either;

-Machine learning that appears to not have an intrinsic bias, but could still miss certain things through the cracks

Or,

-YouTube content overseers with biases, whom determine on their standing what is deemed 'kid friendly' or not

This video right here actually highlights an example of a worst case scenario of content being mistakingly labeled as 'for kids':


GI-tQWj89qU


The implementation of COPPA so far has left much to be desired.

Part of why I was never that keen on this policy in the first place ;D. As I said before, YouTube should've just lobbied Congress for newer regulations that would have made more applicable sense for the modern era we live in.

Despite that though; especially when Machine learning is still years away from being perfected, it's impossible to truly regulate what is and what is not exactly child friendly content. That was the main fear of many content creators, and this obviously inappropriate video being deemed 'for kids' sort of vindicates those fears. It's only a matter of time until YouTube execs begin to have a committee of actual overseers that determine content (like the one I posted above) on its likelihood of getting entangled in an algorithm that a child would likely fall into. But as we've seen with previous 'Adpocalypses', it's not entirely clear which route is rather preferable, especially if you're like me (and others) whom are not that keen on the draconian actions the platform has taken in recent years, either;

-Machine learning that appears to not have an intrinsic bias, but could still miss certain things through the cracks

Or,

-YouTube content overseers with biases, whom determine on their standing what is deemed 'kid friendly' or not


Machine learning can't possibly ever be perfect, because it's all predicated on possibilities or percent confidence. False positives are inevitable. It saw a perhaps thousands of Arthur videos that were very obviously for kids, and so it picked up on Arthur being for kids. Then this one YouTube Poop comes along and maybe it detected that it had some crude language, but all the previous Arthur videos were so overwhelmingly for kids so it probably guessed that there's a 60% chance that the video is targeted at kids and marked it as such. That's where a human has to come in and correct it, and maybe next time it will recognize that any video with "YouTube Poop" in the title is most likely not for kids.

From what I've seen the algorithm is actually very accurate. There were a lot of mistakes at first but that's how machine learning works; it learns and improves as its data sample improves. So if you see a video that's erroneously marked as for kids (or not when it should be), it's better to just report it.

YouTube receives less ad revenue on kid-targetted videos so I doubt they're going out of their way to mark videos as such tbh.

Machine learning can't possibly ever be perfect, because it's all predicated on possibilities or percent confidence. False positives are inevitable. It saw a perhaps thousands of Arthur videos that were very obviously for kids, and so it picked up on Arthur being for kids. Then this one YouTube Poop comes along and maybe it detected that it had some crude language, but all the previous Arthur videos were so overwhelmingly for kids so it probably guessed that there's a 60% chance that the video is targeted at kids and marked it as such. That's where a human has to come in and correct it, and maybe next time it will recognize that any video with "YouTube Poop" in the title is most likely not for kids.

From what I've seen the algorithm is actually very accurate. There were a lot of mistakes at first but that's how machine learning works; it learns and improves as its data sample improves. So if you see a video that's erroneously marked as for kids (or not when it should be), it's better to just report it.

YouTube receives less ad revenue on kid-targetted videos so I doubt they're going out of their way to mark videos as such tbh.


But that's exactly my point. You would have to have a human to fill in the gaps that the algorithm fails to account for. And if there are hundreds, perhaps thousands, of Arthur videos that one could reasonably deduce that most are 'For Kids', then there is no real conceivable way for one human to do that massive amount of quality control. And even if there was a way for a collective unit of humans to find a way to fill in those gaps, we logically would then have to then wonder what exactly are those supposed biases that particular person, or group of people, holds. This adds another layer of complexity in their tasks of being arbiters of content violating COPPA regulations. As seen in the video I just linked above; due to some key words that were mentioned in the video being referenced as why the algorithm is at times massively flawed, not mention that video featuring gameplay in the background, an actual YouTube representative reached out to that YouTuber to mandate that they change the designation of the video from 'General Audiences' to 'For Kids'.

Here is my question towards you Slowpoke, what/who do you trust more in successfully administering this policy? Machine learning or humans? Does your definition of 'success' mean the potential for the policy to never ideally work 100% of the time?

^ That's what's pretty much doomed YouTube. All the old videos around of inactive users, the comment sections are now turned off permanently and you'll never get to interact with people's thoughts and opinions again. It'll just be dead. Comments are a huge part of gaining viewership and subscribers. It's no different than Twitch. You can stream Twitch and all the games you want, but without your chat & interaction, you're absolutely nothing. When you create content or stream content, chat and comments are very important period for success. There are going to be toxicity, I understand, but toxicity exists everywhere in general. It's not like all the comments are toxic. It depends on the specific video and the topic, and what you follow specifically.

Cannot see any end of YouTube now, for with the situation of the world, many a person and organisation are uploading music and activities under the self isolation with the lockdown.

But that's exactly my point. You would have to have a human to fill in the gaps that the algorithm fails to account for. And if there are hundreds, perhaps thousands, of Arthur videos that one could reasonably deduce that most are 'For Kids', then there is no real conceivable way for one human to do that massive amount of quality control. And even if there was a way for a collective unit of humans to find a way to fill in those gaps, we logically would then have to then wonder what exactly are those supposed biases that particular person, or group of people, holds. This adds another layer of complexity in their tasks of being arbiters of content violating COPPA regulations. As seen in the video I just linked above; due to some key words that were mentioned in the video being referenced as why the algorithm is at times massively flawed, not mention that video featuring gameplay in the background, an actual YouTube representative reached out to that YouTuber to mandate that they change the designation of the video from 'General Audiences' to 'For Kids'.

Here is my question towards you Slowpoke, what/who do you trust more in successfully administering this policy? Machine learning or humans? Does your definition of 'success' mean the potential for the policy to never ideally work 100% of the time?


I don't get why you're pitting machine learning against humans. Machine learning doesn't even work without a human helping classify the material it's looking at. The program will do 99.9% of the work and the human will do 0.1%. If you don't report the video for a human to look at then obviously nothing can be done about it.

I will take a 99.9% success rate over a 100% failure rate for sure. I have not noticed a disruption in my viewing habits at all related to this change. The ads are getting longer and worse but that's a separate issue.

Cannot see any end of YouTube now, for with the situation of the world, many a person and organisation are uploading music and activities under the self isolation with the lockdown.


YouTube is here to stay!

A Perfect Example:

As to why COPPA...... is sh!t

Ltw82AC1OZc


How in the world is something like this considered for kids? I used to watch this particular YTP 10 years ago, it's anything but child friendly.


Just goes to show you how YouTube is now a shell of its former self. It's really sad to see after going on the site for several years now.

^ That's what's pretty much doomed YouTube. All the old videos around of inactive users, the comment sections are now turned off permanently and you'll never get to interact with people's thoughts and opinions again. It'll just be dead. Comments are a huge part of gaining viewership and subscribers. It's no different than Twitch. You can stream Twitch and all the games you want, but without your chat & interaction, you're absolutely nothing. When you create content or stream content, chat and comments are very important period for success. There are going to be toxicity, I understand, but toxicity exists everywhere in general. It's not like all the comments are toxic. It depends on the specific video and the topic, and what you follow specifically.


I agree! Many videos that I used to engage with I no longer can do, because of these regulations. Nostalgic commercial breaks from my childhood, funny cartoon scenes, gameplay content of old games I used to be play, even cartoon theme songs, a lot of them have unfortunately blocked the ability for users to comment and save under playlists. And as I said before, this policy is only just getting started, there's going to be wide ranging consequences for these policy changes in the years to come.

In the same way that I don't necessarily trust the YouTube algorithm in determining what is 'kid friendly' and 'not kid friendly', is how I have felt for the last fews years about the platform's efforts in determining 'quality' and 'not quality' content, particularly with news; ie., as many independent news/political commentary channels (Philip Defranco Show, TYT, David Pakman Show, Secular Talk, Democracy Now, Amazing Atheist, Ben Shapiro Show, etc.) having faced immense crackdowns from YouTube in recent years, some smaller channels having to be forced to close down (something as simple as having 'Iran' or 'Covid' in the title of their video is enough for the algorithm to detect the video as being 'against the terms of service', therefor leading to a potential 'strike' on their channel), and even medium sized/big channels that have lost tremendous ad revenue due to these changes. Meanwhile 'established' news sources with channels on YouTube like CNN, CBS, or MSNBC, not to mention nightly comedians from these major networks having clips of their episodes on the platform, have become more showcased in the trending tab for news in recent years, in spite of independent content creators getting the shaft. They're allowed to have 'key words' like 'Iran' or 'Covid' in the titles of their videos, and they still receive many ads on their videos (almost like if you were actually watching TV). Not to mention other valid critiques of this policy over the years.

I've seen how disastrous this policy has been to many aspects and sectors of content on the YouTube platform, and I'm afraid this type of 'censorship' (I would dare call it, 'de-platforming') is only likely to accelerate for years to come. When YouTube is essentially the only video platform in existence, it needs to be treated like a 'public utility' to alleviate that, as such.



I don't get why you're pitting machine learning against humans. Machine learning doesn't even work without a human helping classify the material it's looking at. The program will do 99.9% of the work and the human will do 0.1%. If you don't report the video for a human to look at then obviously nothing can be done about it.

I will take a 99.9% success rate over a 100% failure rate for sure. I have not noticed a disruption in my viewing habits at all related to this change. The ads are getting longer and worse but that's a separate issue.


Slowpoke, I understand the science behind it. What I was asking you before, although I guess you did kinda answer the question but I still want to pick your brain a bit, that based on the science of Machine learning and the biases that YouTube representatives have, do you think that is a fair way of arbitrating 'child friendly content'? The main reason why I ask you this, is because we have had regulations on the entertainment industry in the name of 'protecting the children' for decades now in the United States. I will link to you two examples, the film industry and the video game industry. I'd recommend that you give them a watch:

Ntcs8_QLlxgWv3HDVd22P8


The reason I ask you this is because, for the sheer majority of cases, many of these movie studios and video game studios actively self-regulate themselves. They have committees of people whom evaluate the content and determine on a wide array of contextual standards on what rating that piece of entertainment shall receive. That is essentially what YouTube is doing as well, self-regulation themselves in fear of the government doing it themselves, but here is where YouTube's efforts in self-regulating content 'flips the script' in comparison to other industries in the past. Because there are so many videos on YouTube, you practically need a machine to do much of the heavy load. And once again, while the Machine does do a decent job in doing deductive logic on determining content (categories, to tv shows, to titles of videos on tv shows, to videos on TV shows themselves) it could only do as good as the programmers, whom programmed the functionality of the Machine, could do. In essence, the process of determining 'child friendly content' is thus not a fair arbitration, at least to me, as it was used in other industries in the past. Perhaps if the technology was more advanced and thus more accurate, then maybe you would have a point. But the mere fact that there are still many disingenuous examples of people's content getting misidentified (and thus, getting adversely affected) as 'kid friendly', reveals to me why this way of administering this policy is not great. Watch the two videos to get a better grasp of what I am talking about, and we'll go on from there.