Department of Computational and Data Sciences
Department Seminar
Speaker : Dr. Shivani Kumar, Postdoctoral Research Fellow at the University of Michigan
Title : Computational Social Intelligence for Human-Centered AI
Date & Time : June 15th, 2026 (Monday), 11:00 AM
Venue : # 102, CDS Seminar Hall
ABSTRACT:
Human-centered AI requires systems that can reason not only about language, but also about the social, cultural, and moral dynamics that shape human interaction. In this talk, I present a unified research agenda centered on computational social intelligence: building AI systems that can reason about morality, culture, cooperation, and persuasion within diverse social contexts. Drawing from my PhD and postdoctoral research, I trace a progression from modeling socio-emotional signals in conversations to studying higher-level processes such as moral reasoning, value alignment, and persuasive interaction in human-AI systems. I begin with a brief discussion about my early work on emotions and socially nuanced language phenomena such as sarcasm, where multilingual and multimodal signals help models infer intent, interpret interpersonal dynamics, and provide more socially aware explanations. I then discuss my recent work on morality and culture, including the development of multilingual resources for studying how AI systems perform moral reasoning across languages and communities. Building on these insights, I present ongoing work on persuasion and cooperation in human-AI interaction, investigating how AI systems influence human beliefs, decisions, trust, and collaboration through conversational strategies. Together, these efforts motivate a broader vision for socially intelligent AI: systems that can reason about human values, adapt to diverse cultural and interpersonal contexts, and engage with people in ways that are interpretable, cooperative, and responsible.
BIO:
Shivani Kumar is a Postdoctoral Research Fellow at the University of Michigan working at the intersection of natural language processing, computational social intelligence, and human-AI interaction. Her research focuses on building human-centered AI systems that can reason about social, cultural, and moral dimensions of communication, with particular emphasis on morality, culture, persuasion, and cooperation in language interactions. She earned her PhD from IIIT Delhi, where her work explored conversational AI and socio-affective language understanding, including emotion and sarcasm in multilingual and multimodal dialogues. Her current research investigates how AI systems reason across cultures and languages, how they influence and cooperate through conversation, and how such interactions can be made more transparent, interpretable, and socially aligned.
Host Faculty: Prof. Danish Pruthi
ALL ARE WELCOME
