In this episode, we explore the future of AI in education through a discussion with Lily and David, focusing on the potential and limitations of AI technologies. They break down the future into four key categories: tasks that AI will never achieve perfectly, tasks that are a matter of "when" not "if", transformative innovations on the horizon, and areas with substantial unknowns. The conversation includes examples like the challenges of AI detecting AI-generated content, the potential of AI-based personal tutors, and the complexities of maintaining historical accuracy while promoting diversity. This is the fifth episode in the Responsible AI for Lecturers (RAILect) series on the IDEMS podcast.