Tumblr has also decided that the way to make these subjective calls about what is “art” and what is “adult content” is by using automated tools. D’Onofrio basically admits that these tools don’t work properly, saying in his post that “We’re relying on automated tools to identify adult content and humans to help train and keep our systems in check. We know there will be mistakes”.
That is an understatement. Filters don’t work. We’ve seen this in the copyright context many times. For example, YouTube’s Content ID system works by checking newly uploaded material against a database of copyrighted material and notifying copyright holders if there’s a match. And it resulted in five copyright claims being filed against a video of white noise. Five people claimed they literally owned exclusive rights to static.
Damn good article, imo.