On designing AI systems for explainability and transparency, pitfalls in explainability methods, and transparency in the age of LLMs.