About the Workshop
Recent advancements in AI are enabling conversational user interfaces (CUIs) to communicate with humans at various social capacities, transforming the way humans interact with AI agents through natural conversations. To enhance these conversational interactions, many researchers are turning to a key cognitive-social capability that enables interpersonal communication–Theory of Mind (ToM). ToM refers to humans’ capability of attributing mental states such as intentions, goals, emotions, and beliefs to ourselves and others.
While much prior work covers people’s ToM of CUIs and vice versa, we know little of how these can be intertwined to support the design of human-centered CUIs. Putting together technical and social perspectives of research on ToM in human-CUI interaction, from the CUI’s side and from the human’s side, there is an emerging paradigm that we call “Mutual Theory of Mind (MToM)” in human-CUI interaction. In MToM, we envision both the human and the CUI to possess the capability of ToM and continuously make inferences and attribute mental states to each other during an interaction.
This workshop focuses on exploring three broad areas of ToM in human-CUI interactions:
- Building and measuring CUI’s ToM-like capabilities.
- Understanding and shaping human’s mental state attributions to CUIs.
- Exploring the design and consequences of MToM in human-CUI interactions.
Who Should Attend?
As part of the ToMinHAI workshop series, this ToMinHAI workshop at CUI aims to examine the current practices, challenges, and opportunities in designing, building, and evaluating ToM in human-CUI interactions. To support interdisciplinary discussions, we warmly invite academic and industry researchers and practitioners in disciplines including but not limited to cognitive science, AI, HCI, design, machine learning, robotics, psychology, communication studies, and more to submit work that will inform our understanding of ToM in human-CUI interaction.