links to material from my talk at #DragonCon on security weaknesses in artificial intelligence:

first, a stripped-down set of slides:

https://drive.google.com/open?id=1AuSNm2JQ1CCbvDuHU9xkVhqd_ejE8aYe