Over the past decade, much of the discourse around artificial intelligence has been dominated by performance metrics—accuracy, speed, and scalability. While these dimensions have undoubtedly advanced the field, they risk constraining our collective imagination about what AI can and should do for human lives.

In this keynote, I argue that the HCI community must move beyond performance-centric paradigms to re-envision Human-AI interaction as a means of fostering human reflection, resilience, and well-being. Drawing on my group’s research on extraheric AI and interactive multi-agent systems, I will show how AI systems can be designed not to provide the fastest answer or the most efficient solution, but rather to scaffold deeper thinking, stimulate self-reflection, and support diverse forms of human flourishing.

This requires an integration of insights from psychology, medicine, and the social sciences into HCI design, leading to what I call Well-being Intelligence: an emerging paradigm where AI contributes to mental and social health as much as it does to cognitive productivity. I will also discuss the leadership HCI scholars can take in shaping this transition; they must examine not only usability and adoption but also the ethical, social, and cultural implications of AI design choices. By reframing Human-AI interaction around reflection, resilience, and well-being, I aim to stimulate a broader conversation within the HCI community—and society at large—on how we can design AI that sustains and enriches human life. Read more Read less