AI Meets Social Science: Exploring New Frontiers

What exactly happens when economists and social scientists gather for two days to tackle one big question after another, putting the latest AI tools to the test? The 2025 AI in Social Science Conference, organized for the second year in a row by Anjali Adukia, Assistant Professor at Harris School of Public Policy, and Jens Ludwig, Edwin A. and Betty L. Bergman Distinguished Service Professor at Harris, Director of the University of Chicago Crime Lab, and Co-Director of the Education Lab, and co-hosted by the Becker Friedman Institute for Economics, left attendees with big ideas and even more questions to drive their research and discovery.

Scholars and researchers from leading universities explored how AI will change social science and, conversely, how social science will shape the development and application of AI. They examined new theories, innovative methodologies, and novel datasets with the goal of redefining how AI can push the boundaries of knowledge across disciplines and fields, bringing their insights into conversations about key topics like poverty and education. As the conference came to an end, one major question lingered: With AI frameworks and questions evolving rapidly, how will AI continue to reshape society’s most important questions, challenge assumptions, and inspire real-world progress?

AI’s Role in Policy and Society

Traditional social science topics and conversations such as poverty explored new dimensions and nuances at the conference. Roshni Sahoo, a PhD student at Stanford University, explained what it would actually cost to end extreme poverty through direct cash allocation, offering new ways to quantify and minimize existing levels of prediction error in public policy. Raymond Guiteras, Associate Professor at North Carolina State University, compared traditional poverty alleviation approaches to novel algorithmic targeting and causal inference using text data. He argued this fresh approach could significantly improve interventions, underscoring AI’s vital role in large-scale policy design.

Less conventional themes in social science emerged as well, ones with equally important implications for the potential of AI. Guo Xu, Associate Professor at the University of California, Berkeley, presented an interesting observation, made alongside his co-author Joachim Voth. Their paper, "Legacy on Deck: British Royal Navy" found that sons of naval officers consistently achieved greater success due to inherited traits and selective entry and promotion, revealing patterns of unfair advantage that go beyond established ideas of nepotism. Using AI to predict intergenerational success and skill transmission, the authors highlighted patterns of unfairness that have implications beyond their roles within the navy. On the topic of unfair advantages, additional sessions addressed how facial features and perceptions of trustworthiness and masculinity bias occupational success, demonstrating how data and AI can uncover complex sources of bias and inequality. 

The conference also addressed some of AI’s pitfalls and challenges. Emil Palikot, Assistant Professor of Marketing at Northeastern University and research affiliate at Stanford University, explored how personalized content in educational technology, driven by AI, enhances engagement and retention. But the increased incorporation of user data also raised concerns about data privacy and equity. Anna Lorimer, a PhD Student at the University of Chicago, examined the “AI versus AI” dynamic, where AI-powered content moderation is pitted against AI-driven malicious actors, complicating the regulation of online speech. By assigning numerical and quantifiable values to issues like poverty, nepotism, perception and bias, researchers took the time to showcase unique and practical AI applications in policy, governance, and education. Their insights shed light on the importance of engaging in critical conversations about fairness, bias, and the future societal impact of AI. 

Driving Impact Through Data 

“Across disciplines there is an immense amount of interest in how AI tools are going to impact society, and we have the shared goal of redefining our understanding through new approaches” said Alex Moehring, Assistant Professor at Purdue University. whose research focuses on the economics of digitization.

The conference exemplifies the mission that the Center for Applied AI embodies: to achieve impact at scale by transforming our understanding of AI as it is applied to business and society. Through technical innovation and advancements, alongside gathered understandings of human behavior, conferences like these equip leaders and researchers with the necessary tools and expertise to ensure that AI-driven growth benefits everyone—not just a select few. The rapidly expanding capabilities of AI offer immense promise, and researchers remain committed to harnessing it for broad societal good.

What united the diverse topics in the papers presented? Data. Together with the Becker Friedman Institute for Economics, the Center for Applied AI is proud to help facilitate new ways of discovering and applying data for positive impact. 

Stay engaged with the Center for Applied Artificial Intelligence and follow along on social media as we continue to bridge AI innovation and social science, and collaborate on advancing research for equitable, real-world impact. Connect with frontier research in economics and upcoming events by subscribing to BFI’s Weekly Briefing and follow the Becker Friedman Institute on social media.