As I said, some of these children will make 50 videos with no viewers, but then they create a gymnastics or yoga video that goes viral with predators.
Even if the comments are disabled, these children realize what subject matter grows their subscribers and view count the fastest, so they keep producing more of those videos. Then another kid sees their channel, and notices their yoga video got one million views, so now they try making yoga videos to replicate it. To their surprise, they also gain traction, and the cycle continues.
So, even without a single comment, you still end up with thousands of children making tutorials on how to do the splits, while they cannot begin to comprehend what is taking place and how they're being exploited.
I'm pretty sure the phrase I used was “predatory behavior”, not “comments”, so I'm not sure what your point is.
Strange, because the post bugging the limitation of blocking contents identified other visible manifestations of predatory behavior.
Admittedly, they may be ones that are difficult to distinguish in a single act from acceptable behavior, but there is no reason an automatic detection system needs to consider each action independently in isolation.
edit: holy shit, i just watched the video report. YT in serious trouble.
So instead of trying to ban the videos and restricting user freedom to prevent the videos and deploying automation to recognize them, why don't we ban the predatory behavior and set system rules to prevent and/or direct AI at identifying that.