Stella Liu joins JC explore the complexities of AI trust and evaluation in higher education. They discuss AI’s imperfect nature, the latest trends in open-source agents, and strategies for measuring and deploying AI responsibly. The conversation highlights practical approaches to scenario-based testing, differentiates between frontier models and endpoint solutions, and emphasizes balancing convenience, security, and academic integrity. Gain actionable insights for building rigorous evaluation frameworks and embracing AI as a “perfectly imperfect” tool for next-generation campuses.