SCHEDULE 2026

keynote: “Artificial agents, real relationships”

May 13, 2026 | 9:35 AM – 10:30 AM

Speaker: Kashmir Hill, The New York Times

AI systems are increasingly influential in our day-to-day lives. From recommendation engines that shape our attention to chatbots that offer companionship, advice, and emotional support, artificial systems are becoming privy to sensitive information about us and steering our decision-making. But what happens when software begins to function as a "someone" rather than a "something"? In this opening keynote, Kashmir Hill, investigative journalist at The New York Times, draws on her distinctive reporting at the intersection of technology, power, and lived experience to explore how AI systems influence belief, mediate relationships, and quietly reshape norms around privacy, persuasion, and trust.

AI as social agent: Intimacy, empathy, and persuasion

May 13, 2026 | 10:45 AM – 12:15 PM

Speakers: Thomas Costello, Carnegie Mellon University; Anat Perry, Hebrew University of Jerusalem; Kashmir Hill, The New York Times (respondent)

Moderator: Michael Inzlicht, University of Toronto

AI systems are increasingly functioning as social agents. People form deep attachments to AI companions, seek emotional support from chatbots, and sometimes revise long-held beliefs through AI conversations. This session explores a fundamental question: What makes artificial agents effective at both emotional connection and persuasion? Through journalism and experimental research, our speakers examine how AI's capacity for personalized, patient interaction shapes human relationships and beliefs. We investigate what people value in AI empathy versus human empathy, how AI conversations can shift conspiracy beliefs, and what real-world stories reveal about both the benefits and risks of AI companionship. These cases illuminate broader questions about authenticity, trust, and the boundaries between simulated and genuine human connection.

Benchmarking minds: How to evaluate cognitive capacities in general-purpose AI?

May 13, 2026 | 1:15 PM – 2:45 PM

Speakers: José Hernández-Orallo, University of Cambridge (virtual); Raphaël Millière, University of Oxford

Moderator: Karina Vold, University of Toronto

As AI systems increase in versatility and scope, assessing their potential cognitive capacities becomes urgent. This panel brings together researchers from AI, philosophy, and cognitive science to explore how we can meaningfully evaluate such systems. We’ll discuss what constitutes a “cognitive capacity” or an “inner life”? How can approaches from cognitive science, comparative psychology, and psychometrics translate—or fail to—when applied to general-purpose AI? What are the challenges of applying traditional benchmarks to identify emergent cognitive properties? What are the implications of these issues for safety, trust, and alignment?

AI and Misinformation

May 13, 2026 | 3:00 PM – 4:00 PM

Speaker: Jevin D. West, University of Washington

Moderator: David Lie, University of Toronto

Session description coming soon.