Some comments/ideas/takeaways: 1) According to this video, it is very easy for some troll living in another country, who has a bot network, to extort money fro

Some comments/ideas/takeaways: 1) According to this video, it is very easy for some troll living in another country, who has a bot network, to extort money from YouTube creators by gaming YouTube's takedown request system: "I'll get your video taken down unless you pay me" (Sounds a whole lot like patent trolls -- but applied to videos). 2) YouTube (or any other video sharing site, present or future), should conduct a thorough investigation as to how their system for user requested takedowns -- could be gamed by bot networks / multiple phony user accounts requesting takedowns -- and what could be done about it, using what happened here as a baseline for that investigation. Note that this also applies to non-video sharing sites where user takedowns can be requested, i.e., craigslist -- if that process is automated in any way, it should be thought about from the perspective of "how do we protect against swarms of automated bots / fake user accounts, requesting takedowns?" 3) Any ML/AI algorithm that is used in the takedown determination process should consider the reputation of the user whose videos are being requested to be taken down. This includes such things as 'how long has the user been a user', how many other videos have they uploaded, how many likes they've received over the years, how many video views, and how many real legitimate complaints, verified by an actual YouTube human staffer -- have occurred on that account over the years (feeding that data in only after human YouTube staffer review, etc.) 4) It wouldn't be a bad idea (if YouTube is short of staff), to create another tier of users, "review users", that is, these users would be / could be called upon to review content that was flagged for removal/deletion for one reason or another. If enough "review users" (randomly selected) collectively vote that the content should be removed, THEN and only then it gets queued for a human being staffer at YouTube for final review... otherwise it's not banned, and sticks around... but this would hopefully save a lot of work for human YouTube staffers -- much in the same way automated voice assistants do for human telephone operators or receptionists... any AI/ML algorithms could be the gatekeeper for this tier of users; if the AI/ML has a problem with the content -- escalate it to this tier of users for further review... 5) To implement #4, User accounts, agreements, and YouTube's suggestion algorithm -- would have to be modified (basically you're adding another user account type and another agreement type) to support the functionality. 6) No matter how great YouTube is, it's a human created system like all others, and as such, subject to failure. As the old expression, "Don't put all of your eggs in one basket" goes, similarly: Don't host all of your videos on one and only one video hosting site. Just like you wouldn't store your important files in one location only, without making/having backups...