Cybersecurity, like other industries, has seen an explosion in the use of artificial intelligence (AI) and machine learning (ML) technologies in recent years. AI and ML can help to automate tasks. Data-driven approaches in general can draw patterns from vast volumes of data far quicker than humans are can. This episode summarises the state of AI for security at the time of writing and highlights some of the considerations to guide whether it is an appropriate approach for a given problem, common pitfalls to avoid, and human-AI ecosystems.AI is challenged by several open research areas including lack of transparency, robustness to concept drift, and the security of AI systems themselves. This topic guide is for those looking to build and/or procure AI solutions to use for cybersecurity applications. Some sections are more relevant for those building and others to those procuring solutions. We speak with CyBOK AI for Security author Matilda Rhode for an overview of the topic.