We are seeing a new approach to security that is rippling across network defenders, products, and attackers alike. The approach is based on the idea that you can improve security on data by harnessing data to improve security. This requires transitioning from appliances that shrink data volumes to cloud approaches that capture more data than ever before. Innovative organizations and security vendors are seeking signals across user, device, and application activity, and building learning systems to master security insights from them. In this talk, John Lambert will discuss these trends framed through specific exploit examples, highlight new risks that arise and provide insights on how attackers are adapting to these data driven systems. Attendees will walk away with a better understanding of these trends, the privacy implications of these security learning systems and how to think about augmenting them with additional signals or applications.