This course is a comprehensive course that explores the challenges posed by Large Language Models (LLMs) when they exhibit excessive agency—taking unintended actions or making decisions beyond their designed scope, which can lead to misalignment with human values and unintended consequences.
Through theoretical insights and practical case studies, participants will learn to detect signs of excessive agency, understand the associated risks, and implement effective strategies to mitigate these risks, ensuring AI systems remain aligned with ethical standards and societal values.