This paper delves into the unique security challenges posed by Generative AI, and outlines potential research directions for managing these risks.