UMSI convenes scholars and industry leaders to shape human-centered AI

As artificial intelligence reshapes nearly every aspect of society, the University of Michigan School of Information and the Michigan Institute for Data and AI in Society (MIDAS) is convening leading voices from academia and industry to chart a more human-centered future for the technology.

The U-M Symposium on Human-Centered AI will take place Wednesday, Oct. 29, at Rackham Amphitheatre. The event will feature keynote addresses, research talks, poster sessions and a panel discussion, all centered on designing AI systems that amplify human capabilities while addressing issues of fairness, transparency and trust.

“The School of Information was invented to bring together different disciplines and ways of understanding the world,” says Andrea Forte, dean of UMSI. “Today, we need this university-wide breadth to ensure that we don’t just build better technologies, but technologies that truly improve people’s lives.” 

By centering people in conversations about the future of AI, the event highlights U-M’s interdisciplinary strength and its commitment to shaping technologies that work for society as a whole. UMSI’s strength as a cross-disciplinary hub and MIDAS’ role in fostering data science innovation combine to create an environment where ethical, social and technical considerations intersect.

“UMSI faculty and students have strong intuitions about the social and human impact of technologies. It’s in the water here,” Forte notes. “As the world looks for voices that can address not only what AI can do, but also its long-term and far-reaching effects, the work of UMSI scholars will become increasingly vital.”

The symposium brings together a distinguished lineup of keynote speakers and featured presenters, each offering unique perspectives on how AI can be designed with people at the center. They include: 

• Elena L. Glassman, a computer science professor at Harvard University. She is a leading researcher in human-computer interaction, focusing on AI-resilient interfaces that make complex systems more accessible and transparent.
• Margaret Mitchell, chief ethics scientist at Hugging Face and co-founder of Google Ethical AI. Mitchell is internationally recognized for her pioneering work at the intersection of machine learning, ethics and inclusion.
• Daniel S. Weld, chief scientist at the Allen Institute for AI and professor emeritus at the University of Washington. Weld brings decades of expertise in AI systems, human-computer interaction and responsible innovation.
• Diyi Yang, assistant professor of computer science at Stanford University. Yang is a rising leader in human-centered AI. Her research investigates the social aspects of language and builds socially aware natural language systems to enhance human communication and collaboration.
• Lauren E. Gillespie, assistant professor at U-M’s School for Environment and Sustainability. Gillespie develops AI-integrated approaches for large-scale ecosystem monitoring, using foundation models to improve biodiversity forecasting and conservation decision-making.
• Q. Vera Liao, associate professor of computer science and engineering at U-M. Liao is a specialist in human-AI interaction, explainable AI and responsible AI, with award-winning research connecting the HCI and AI communities.
• Abigail Jacobs, assistant professor of information at UMSI and an assistant professor of complex systems at U-M’s College of Literature, Science, and the Arts. Jacobs is an expert in trustworthy AI and algorithmic bias. A 2025 Microsoft Research AI & Society Fellow, Jacobs’ current research focuses on measurement and validity as a lens for governance in responsible AI, structure and inequality in sociotechnical systems and social networks.
• Huteng Dai, assistant professor of computational linguistics and phonology at U-M’s the College of Literature, Science, and the Arts. Dai is a computational linguist and cognitive scientist. His research focuses on understanding how humans learn sound patterns from noisy, real-world data. He builds interpretable learning models that are not only mathematically well-defined but also succeed on real-world corpora.

Two of the symposium’s speakers, Dan S. Weld and Diyi Yang, bring distinct but complementary perspectives to the future of human-centered AI. For Weld, the timeliness of this theme lies in AI’s growing technical sophistication and the need to make these systems trustworthy collaborators.

“Modern AI systems are rapidly gaining impressive capabilities, ranging from reasoning to tool use,” Weld says. “In order for people to best harness these skills, it’s important for us to design AI systems that are also good collaborators.” 

Yang, meanwhile, is looking ahead to how human-centered AI can expand what people are able to do in everyday life. 

“I’m most excited about building AI systems that truly augment human capabilities across research, work and well-being,” Yang says. “Two areas where I see huge potential are education and healthcare, where I’m especially excited about how AI can enhance the way we learn, teach, and care for others.”

UMSI associate professor and associate professor of electrical engineering and computer science at U-M’s College of Engineering and MIDAS affiliate David Jurgens underscores why these conversations matter now more than ever. He is co-organizing the symposium with Sabina Tomkins, UMSI assistant professor and a faculty associate at U-M’s Center for Political Studies. 

“AI is touching nearly every part of society, which raises important questions about how we center humans in the development of AI,” Jurgens says. “I hope attendees leave with a clearer sense of how human-centered approaches can be applied in their own research, teaching, or applications — and with new collaborations that advance this vision of building AI that supports people and society.”

Event Details: 

Symposium on Human-Centered AI

Date: October 29, 2025, with an invitation-only workshop on October 30, 2025. 

Location: Rackham Amphitheatre, 915 E. Washington St., Ann Arbor, MI

Hosted by the Michigan Institute for Data Science & University of Michigan School of Information