I would like to create my own forum, but concerned about the possibility of users uploading harmful images. Does anyone have any experience implementing an automated system that detects and blocks harmful content or running a user generated content website in general?
the biggest advice I'd give is to be proactive rather than reactive, screenshots last forever. In terms of proactivity, the easiest way is user vetting, the hardest is real moderators. The second biggest piece of advice is to emphasize concentration via invisible limitation, people are less willing to post harmful things when they're with people who know them, when their reputation actually matters. Image moderation is also quite hard. Notice the level of civility on HN, now notice what feature it lacks. Images create asymmetry, in text, we're all equal