As artificial intelligence has boomed over the past several years, industry commitment to responsible AI has often failed to keep pace, perhaps because executives do not see such practices as essential to their bottom line. This fall, iSchool professor Min Kyung Lee will embark on an investigation, funded by a coalition of major philanthropies including the Ford Foundation and the Omidyar Network, to identify the relationship between companies’ responsible AI practices and their financial outcomes.
“The goal is to understand whether responsible AI-related practices, especially in generative AI, have any impact on financial materiality, meaning whether there is a profit or how well they're doing in the stock market,” Lee says.
The project is a collaboration with scholars at the Kellogg School of Management at Northwestern University, who will take the lead in assessing financial performance across various metrics. Lee’s team at the iSchool, meanwhile, is developing a framework to index firms’ investments in responsible AI, drawing on news reports and 10-K financial disclosures.
Lee expects to begin the process by drawing on existing attempts to define responsible AI best practices. “There's a lot of work from standard-setting organizations,” Lee says. “The National Institute of Standards and Technology has an AI risk framework, the European Union AI Act also lays out important dimensions, and the International Labor Organization has their own version. Right now, we are trying to survey these existing responsible-AI-related frameworks, then think about how we synthesize them for indexes and measures. We are also trying to see how we can use NLP-like AI to extract this information from either news reports or firms’ disclosure materials.”
Lee’s team on the grant will include iSchool PhD student Angie Zhang, iSchool undergraduate Naomi H. Ichiriu, and two UT students from outside the iSchool: computer science PhD student Juan Diego Rodriguez and biomedical engineering and business and public policy undergraduate Chibudom Okereke. The grant covers research expenses of $150,000 over one year.
This project represents a new frontier in Lee’s research. She says she was motivated by the funder’s call for projects that put responsible AI front and center.
“Even though it’s a new kind of work for me, I responded to this funding call because I think that looking at responsible AI’s impact on financial performance is important as a lever to increase these practices,” she says.