Have you ever tried to build a system to categorize text, images, or video as “offensive”? What about combinations of those three? With “memes” today, they have both images and text. Video can have all three. Somehow you are expected to grab a text scanning engine, have a way to analyze images for “adult” content, and scan all of that in a video stream. The problem has been around a while, and many have tried to solve it. Most of the solutions so far have been slamming together various bits of software packages and a lot of work to get them all to work together. Come to my talk and let me show you how the Cognitive Services team at Microsoft have created a first class system to attack that problem with code, machine learning, and the help of the almighty cloud. Not only will I demonstrate how to programmatically configure and control the content, but also how to include a human in the process for the more difficult decisions. I will have code demonstrations and a ton of fun talking about how to deal with folks like myself that sometime cross the “line”.