I wonder how far we are from automating this with image recognition. The interesting thing is policy aspect to the learning problem, does an post/image/video vi

I wonder how far we are from automating this with image recognition. The interesting thing is policy aspect to the learning problem, does an post/image/video violate a collective set of rules? They are certainly building a huge enough labelled dataset of images, decisions, and policies. I wouldn't be surprised if this went automated in 5-10 years!