Employing machine learning for cybersecurity offers tremendous benefits and great promise. However, there are also many risks that security professionals must understand. This presentation seeks to demystify machine learning for the non-data scientist so that cyber security professionals can harness its full potential. This presentation will briefly cover some underlying machine learning concepts and how machine learning can improve security. Then, we will look at the possible vulnerabilities and risks, including attacks on the machine learning process, data, and models. Finally, the presentation will explore implementing adversarial machine learning to test and protect our models and data. This presentation focuses on using machine learning for cybersecurity; however, the concepts could apply to employing machine learning in any adversarial environment.