Explainable AI for Physical Layer Security in Next-Generation Wireless Networks

Physical layer security (PLS) has garnered increasing attention as a complementary solution to conventional crypto-graphic techniques for addressing the diversity of use cases, deployment scenarios, and device capabilities in today's wireless networks. Integrating artificial intelligence (AI) into PLS offers a promising solution to various multi-dimensional, heterogeneous, and complex challenges stemming from the growing complexity of networks. However, the opaque nature of AI has raised concerns regarding its trustworthiness and interpretability. This paper addresses these concerns by emphasizing the crucial role of explainability in AI-based PLS in several use cases of next-generation networks. A particular focus is placed on physical layer authentication through an illustrative case study, aiming to enhance AI-empowered PLS's practicality and trustworthiness. The paper concludes with a discussion of some critical future research directions.