> yet you would punish them if they had repeat offenders commenting under their videos
You misunderstand. If the uploader keeps uploading videos of minors that garner mass sexual comments, the uploader gets banned.
Uploader uploads preteen girl deepthroating popsicle. Lots of pedophiles come along and express their pleasure in the comments with water squirting emoticons and such. Bam, video flagged and taken down. Uploader uploads preteen daughter doing handstands/nipslips. Pedophiles return to comments to express pleasure in their sick way. Uploader is now repeat offender and account is banned.
I don't see why this is a big problem. If it's a family video make it private.
ML can be used to train algo on the latest slang.
> I have no doubt Google/YouTube/law enforcement currently have the capabilities of tracking down anybody using these services without any paper trails, that's in no way part of the issue.
Really? Why aren't they doing it? How do they track down people uploading via VPNs and such? Tying account to bank makes it much more difficult to evade law enforcement.
> You misunderstand. If the uploader keeps uploading videos of minors that garner mass sexual comments, the uploader gets banned.
I understood you perfectly fine, that's why I said this approach is easy to exploit by simply spamming the videos of some uploader you don't like with "sexual comments".
The qualifier of "uploads videos of minors" does not work because there is no kind of ML/AI that could reliably recognition children and their ages, and whether what they are doing is "sexualized" or not.
For that, you would need human moderation, and that's not just impractical, it's simply impossible because during one second more than 1 hour of content is uploaded [0]. People are uploading content faster to YouTube than any reasonable human workforce could ever hope to review, without generating a steadily growing massive backlog.
> I understood you perfectly fine, that's why I said this approach is easy to exploit by simply spamming the videos of some uploader you don't like with "sexual comments".
You'd need to have dozens of puppet accounts. Obviously you could take into consideration whether all the sexual comments are coming from one person/ip address.
You misunderstand. If the uploader keeps uploading videos of minors that garner mass sexual comments, the uploader gets banned.
Uploader uploads preteen girl deepthroating popsicle. Lots of pedophiles come along and express their pleasure in the comments with water squirting emoticons and such. Bam, video flagged and taken down. Uploader uploads preteen daughter doing handstands/nipslips. Pedophiles return to comments to express pleasure in their sick way. Uploader is now repeat offender and account is banned.
I don't see why this is a big problem. If it's a family video make it private.
ML can be used to train algo on the latest slang.
> I have no doubt Google/YouTube/law enforcement currently have the capabilities of tracking down anybody using these services without any paper trails, that's in no way part of the issue.
Really? Why aren't they doing it? How do they track down people uploading via VPNs and such? Tying account to bank makes it much more difficult to evade law enforcement.