Jenn Wortman Vaughan - Can Generative AI Deepen Our Own Thinking? Supporting Appropriate Reliance, Human Agency & Beyond

Human-centered AI that empowers rather than overrides The rise of generative AI has amplified existing sociotechnical challenges in responsible AI and introduced new ones. When we come to depend on AI systems, we risk overreliance, loss of agency, diminished innovation, atrophied skills, and fewer opportunities for joy and self-fulfillment in our work. In this talk, I argue we should work towards a different vision: a future in which AI helps people grow and flourish—enriching human capabilities, deepening critical thought, and safeguarding our agency. I will share some very tiny steps towards this very lofty goal, drawing on recent empirical work, including controlled experiments on fostering appropriate reliance and a qualitative study of designing LLM-based writing-support tools to uphold agency in science journalism.