Abstract: Terms such as “Machine Learning” and “Cybersecurity” are everywhere, today. Increasingly more operational systems are integrating machine learning (ML) methods (sometimes for cybersecurity). Such additional components, despite providing advantages, also expand the threat surface that can be exploited by attackers to compromise a given system. Abundant research has studied the relationships of ML with cybersecurity. However, from a practical viewpoint, scientific findings tend to be overlooked; and research itself can sometimes overlook some practical aspects of operational cybersecurity. Regardless, as ML becomes more mature, it can be easily used as a weapon to carry out malicious deeds. In this talk, I will explore some practical scenarios entailing the interplay between ML and cybersecurity, highlighting several blind spots and open issues.