I have friends who work on teams in Facebook who use ML to detect and filter out “undesirable content”. It’s a difficult problem but they’re a pretty intense team (workaholics) so I wouldn’t say they spend 0 effort.