We research, design, and build AI-powered technologies that account for humans' complex social dynamics and empower users in their situated ecosystems (e.g., professional domains, healthcare systems, and social networks). Here are some of the key projects that we work on at the moment:

1. AI-mediated social support and its implications for human social interaction: From conflict resolution and intergenerational communication to relationship advice, people increasingly turn to AI to read between the lines and navigate all sorts of interpersonal situations. We examine whether, when, and how people adopt AI’s suggestions and what downstream impacts those choices have on their real-world social interactions and relationships.
2. Companion AI that connects users back to their support networks: Companionship is now one of the most common uses of generative AI. How do such AI uses become detrimental, and when do human interventions become necessary? Ultimately, how do we disengage users from harmful reliance on AI and reconnect them to existing human support within their personal and healthcare networks?
3. AI teammates that support positive group dynamics: AI now holds the agency not only to facilitate but also to participate in teams and groups. How do AI agents steer social interactions in groups and teams? How do we design agents that are more socially aware and support positive dynamics?
4. Collaborative AI tools that foreground human expertise: AI assistance at work holds unprecedented potential to democratize expertise across domains. How do we design AI tools that empower users while spotlighting the value of human work? How should workflows and AI use policies be framed to support this goal?

By keywords, here are some of the subjects that best capture the lab's interests:
#   human-AI collaboration
#   human-agent teamwork
#   societal impact of AI + scalable methods for HAII research
#   digital mental health
#   future of work

The lab is grateful for support from Google Research, Amazon Science, OORI, Slack, USC Annenberg, and USC Center on Generative AI & Society.

08/2026   Two papers accepted at EMNLP! Check out our latest work on human-AI collaboration and AI companionship in vulnerable conversations.

07/2026   Two papers accepted at AIES! Check out our latest work on Sociotechnical Outcome Taxonomies (SOT) and opem model marketplace.

07/2026   We are grateful to receive support from Princeton CITP and Harvard Berkman Klein Center to expand our work on AI's impact on social interaction.

07/2026   We ran our first 2-day workshop with youth of color (YOC) to co-design AI for mental health and wellbeing support 🩵

06/2026   Our reporting toolkit for LLMs in behavioural science is out in Nature Human Behaviour.

06/2026   John successfully defended his dissertation and is now Dr. Bunyi 🎓 We are so inspired by his work (see more here and here) on how people's support systems influence their decisions to adopt digital mental health technologies!

06/2026   We received Top Methodology Paper Award at ICA 2026 🏆

04/2026   We received Best Paper Honorable Mentions (top 5%) at CHI 2026 🏆 Check out our latest work on digital companionship, disclosure of AI use in freelance work , and a responsible design framework for AI-mediated mental health and wellbeing support.

02/2026   CounselBench is accepted at ICLR for 🌟Oral Presentation🌟 (top 1-5% of all submissions) 🏆 Shout-out to Yahan Li as the student lead of this project!

02/2026   Thrilled to receive a grant from the OORI Research Catalyst Program to support our work on democratizing LLM for algorithmic decision-making! Tripple congrats to collaborators, Phebe Vayanos, Jieyu Zhao, and Peter Phung!

01/2026   So honored to receive Best Paper Award at ICA this year! Shout-out to Chang Wan as student lead of the paper 🏆

10/2025   Honored to receive Google's Academic Research Award!! Along with Munmun De Choudhury's team, we will work on idnetifying criticial cues that signals harmful interactions between users and AI companions and needs for human interventions.