There still seem to be some many basics that FB and YouTube get wrong with their "automatic" video moderation. Use people's "report" button clicks as an early

There still seem to be some many basics that FB and YouTube get wrong with their "automatic" video moderation. Use people's "report" button clicks as an early indicator that something should be looked at - a big flag. Look at certain words or phrases that are either unambiguously bad or possibly like "kill" or "why I hate ..." - a medium flag Phrases that are less certain, = a small flag. Block some things proactively, moderate others, react to the others as they are reported. I don't think people are bothered that bad stuff gets posted as much as these companies not doing anything about it when it is reported. If 1000 people flag a video, up to the top of the list. If the flagging is malicious, then downgrade the reputation of the accounts that flagged it. If new accounts upload stuff, rate-limit it in some way etc. I know I am making it sound really easy but with how ever many 1000 Developers, it shouldn't be impossible for someone like FB to do this much better.