GroupLens Research Team Earns Best Paper Honorable Mention at CHI26

Department of Computer Science & Engineering (CS&E) PhD graduate Mo Houtti was the lead author on the paper that earned Best Paper Honorable Mention at the CHI 2026 Conference in Barcelona, Spain. CHI, the ACM Conference on Computer-Human Interaction, is the premier international conference in the field of Human-Computer Interaction. The paper titled, “Opportunities and Barriers for AI Feedback on Meeting Inclusion in Socioorganizational Teams,” included seven researchers from the GroupLens Research Lab at the University of Minnesota - Houtti, Moyan Zhou, Daniel Runningen, Surabhi Sunil, Assistant Professor Harmanpreet Kaur, Professor and CS&E Department Head Loren Terveen, and Assistant Professor Stevie Chancellor - as well as Leor Porat from the University of Illinois Urbana-Champaign.

The paper highlighted the third phase of Houtti’s PhD project which explores workplace bias in video conferences and interventions to mitigate bias. Partnering with Cisco Research, the research team built upon theory from human-computer interaction and social psychology to find a new way to use AI to improve meeting inclusion in real-world settings. They created an AI chatbot that would help provide feedback to users following group meetings. This chatbot was first tested in a lab experiment and was later refined and launched in a field study with a small consulting firm based in New York City and Delhi.

While the chatbot performed very well in the controlled lab setting, there were some adjustments that needed to be made after implementing the technology in a real workplace. Workplace culture and personal relationships had a major impact on how the chatbot was used and how feedback was received.

“We spent a lot of time focusing on feedback mechanisms to improve group performance,” Houtti said. “Feedback can be difficult for people to exchange in any setting, but especially in a corporate environment with hierarchies. We found that people find it easier to give and accept feedback to an AI agent than face-to-face with colleagues. Conversely, people are also less likely to listen to AI agents when they give difficult feedback. It was easier for people to take in the information, but also easier for them to ignore it.”

In order to account for some of the workplace dynamics influencing the impact of the AI chatbot, the team introduced the “induced hypocrisy procedure” to get more people to engage with the chatbot and adjust their behavior in meetings. The Induced hypocrisy procedure works by asking someone to agree with a norm - for example, recycling is good for the environment. Then you have a person reflect on a time that they did not behave in accordance with that norm. When a person goes through those two steps, they want to behave in accordance with that norm going forward to avoid feeling like a hypocrite. Ultimately, some of the realities of day-to-day work played into the overall effectiveness of the chatbot. 

“In the field study, there were some clear benefits to these systems and it caused people to slow down more and reflect on their presence in meetings,” Houtti said. “However, in the context of a fast-moving consulting firm, there is not always time to slow down and reflect. We need to find ways to incorporate this process in a way that can more easily integrate into the workflow of a company.”

This Best Paper Honorable Mention is the second award Houtti and team have earned for this project. They also earned the Honorable Mention Award at the 26th ACM Conference on Computer-Supported Cooperative Work and Social Computing (ACM-CSCW) for their paper titled, “‘All of the White People Went First’: How Video Conferencing Consolidates Control and Exacerbates Workplace Bias”. That paper also received an Impact Recognition for strong examples of work that demonstrate clear potential for real-world or practical impact.

“When I first took on this project, I told my advisors that I really wanted to do something ambitious, like building a system that was robust enough to be deployed in the real world. I’m happy that we put in that work and that it was recognized. The hard work definitely paid off. This is one of the papers that I am proudest of in my PhD journey and it is definitely my most collaborative paper. I think this award is directly attributable to the fact that we had so many great people working together on this.”

Co-advised by Chancellor and Terveen, Houtti finished his PhD in August 2025, and currently works at Microsoft as a senior applied scientist. Learn more about this work in the published paper.