Explore the concept of explainability in speech processing through this comprehensive lecture from JSALT 2023. Delivered by Jean Francois Bonastre, the talk delves into the journey from theoretical aspirations to practical applications in the field. Learn about the challenges and advancements in making speech processing systems more transparent and interpretable. Gain insights into the latest research and methodologies presented at the 30th edition of the JSALT workshop, co-hosted by LIUM (Laboratoire Informatique de l'Université du Mans) and Allo-Media in Le Mans, France. This 1-hour and 27-minute presentation offers valuable knowledge for researchers, practitioners, and students interested in the intersection of explainable AI and speech technology.