AI is steadily flooding our world sipping into various verticals from Autonomous cars, Robots to Defense, Media and Smart homes. Hundreds of new startups are implementing Ai solutions worldwide and we are getting closer to the point where machine learning-based solutions will eat traditional algorithms. While we more and less understand how to deal with software vulnerabilities we have no clue what’s happening in ML-based solutions and how it's possible to hack them except probably adversarial examples which recently attracted media attention after 5 years from the initial invention. But this area is rapidly growing and we are getting to the point when there will be over 2000 research papers presented on this topic and some of them will hit media only in 5 years or so. In this presentation I will show what is happening in AI security industry, the most closed cybersecurity area. We will discuss the most critical AI applications such as face recognition, self-driving cars, voice assistants and their latest attacks. Then we will discuss ML algorithms such as classification, regression, reinforcement learning, clustering, etc… and how to attack them. An finally we will look at the particular attack methods such as adversarial, privacy, poisoning, backdoor, reprogramming, and how they evolving.