2025 Research Day Highlights Growth with Global Impact
Earlier this month, the Data Science Institute (DSI) held its third annual Research Day to celebrate achievements from the past year. Sessions demonstrated the breadth of the DSI’s work in fields from internet access to biology, and showcased partnerships across campus and beyond–from the University of Chicago Law School and the Harris School of Public Policy, to the Illinois Office of Broadband and national governments.
The day drew members of the UChicago community as well as industry partners. “Today’s audience and agenda really underscore the DSI’s dual role as a hub for interdisciplinary research on campus and a meeting point for collaborators using AI to drive innovation beyond academia,” DSI Faculty Co-Director and Morton D. Hull Distinguished Service Professor Michael Franklin said in his opening remarks.
Read on for highlights touching on data privacy, democratizing AI and internet access, and more.
From Revolutionizing Weather Forecasts to Improving Data Ecosystems
The DSI’s newest initiatives, AI for Climate (AICE), Complementary AI, and Data Ecology, shared their accomplishments since launching last spring.
From the AI for Climate Research Initiative (AICE), Faculty Directors Pedram Hassanzadeh, Associate Professor of Geophysical Sciences, and Tiffany Shaw, Professor of Geophysical Sciences explained how AI has precipitated a revolution in weather forecasting by making weather simulations faster, lower cost, and more accurate than ever, enabling earlier event prediction and expanding access to millions globally.
As one example, in the past year, the AICE team has partnered with the Indian government to share AI-supported forecasts with 50 million Indian farmers on a weekly basis via text message. In an environment where monsoons can upend a season’s worth of cultivation (and livelihood), better forecasts could empower farmers to make cultivation choices that work with the weather rather than against it. Through a collaboration with the Gates Foundation and the government of the United Arab Emirates, the AICE team plans to further democratize forecast access over the next three years by training government forecasters across 30 low- and middle-income countries.
Faculty Directors of the Data Ecology Initiative Assistant Professor of Computer Science Raul Castro Fernandez and Assistant Professor of Law Bridget Fahey shared how in an increasingly data-driven world, rather than simply patching problems as they arise, their team is studying how to design data systems to maximize data’s value while controlling its use and impact. As just one illustration of why this is needed, the team pointed to the informal market for personal data between governmental organizations. Professor Fahey published a recent article in The Atlantic on the risks of such unregulated exchange.
Fahey and Fernandez propose as an alternative a system that collects, protects, and tracks queries of data. Professor Fernandez has already developed the blueprint for such an alternative—the data escrow—which incentivizes data quality and lets stakeholders follow how data is used.
Associate Professor of Computer Science and Data Science Chenhao Tan and Professor of Sociology and Director of the Knowledge Lab James Evans shared the latest concepts from the Complementary AI Initiative, including the challenges of leveraging AI to complement human intelligence when users reward AI for alignment with their own values.
Representing the Internet Innovation Initiative (III), Faculty Director and Neubauer Professor of Computer Science and the College Nick Feamster recapped his team’s work testing the limits of internet speed measurement to facilitate real-time solutions. In addition, the III collaborated with the federal Broadband Equity, Access, and Deployment (BEAD) Program to extend access to high-speed internet to 175,000 homes, businesses, and schools across Illinois.
Next, from the AI+Science Initiative, the Schmidt AI in Science Fellows presented their research.
The day also featured presentations by DSI faculty including Victor Veitch, on how semantic meaning is encoded for large language models (LLMs); Tian Li, who discussed applying human preferences when scaling machine learning (ML) to modern systems; Haifeng Xu, on his work with Instagram and Meta to diversify recommendation systems and disincentivize clickbait; Nikos Ignatiadis, on leveraging ML to share data and control bias; Ce Zhang, who shared measures his team at Together.ai is taking to democratize intelligence by making it faster, simpler, and lower cost to use; and Alex Kale, who shared research on using data visualizations as a decision aid to support participatory budgeting in Chicago.
The day concluded with a reception and poster session where researchers and attendees could take in research by DSI postdocs and continue the conversation.
“The DSI continues to be a catalyst for groundbreaking, cross-campus research at the intersection of foundational AI and real-world impact,” said Executive Director of the DSI David Uminsky. “This year, with the launch of new initiatives and the arrival of exceptional faculty, our footprint has expanded into critical domains such as climate science, law, and economics. It’s an exciting moment for the DSI, and we are just beginning to glimpse what’s possible when world-class researchers converge to tackle society’s grand challenges through data.”
Thank you to everyone who attended for their engagement at Research Day, and congratulations to this year’s presenters for all their work throughout the year! We look forward to making new discoveries together in the new year.
