LLMs for XAI: Future Directions for Explaining Explanations