The Data Behind Agents: Building Intelligent, Safe, AI Systems

AI agents dominate current industry discussion, yet clear explanations of what they are, why they fail, and what roles they are creating remain scarce. This session offers a foundational introduction for students and early-career technologists. We’ll examine what distinguishes an agent from a conventional chatbot — tools, memory, and planning loops — and trace an agent through the completion of a realistic task. We then turn to the less-demonstrated half of the problem: the data an agent requires to be useful, and the boundaries it requires to be safe.