How should humanistic scholars evaluate and harness the potential of generative AI? In what ways can they contribute to its development?

An interdisciplinary group of researchers is convening at the Neubauer Collegium for Culture and Society to answer these questions, which are growing increasingly urgent as AI technologies become embedded in every field of study.

The Humanistic AI project, headed by UChicago professors Hoyt Long and Chris Kennedy, aims to identify the opportunities and challenges that generative models present across a wide range of disciplines—including literature, linguistics, philosophy, sociology, computer science and more—and to articulate a strategic vision for how the humanities, humanistic social sciences, and computer sciences can collaborate to advance AI research.

The team is also investigating the impact of generative models beyond academia, notably on creative processes that use AI tools. Participants will develop a set of case studies to lay the foundations for a new approach to humanistic research and cultural activity that accounts for these rapidly evolving technologies.

“I could not be more excited about the ways in which our faculty in the arts and humanities are thinking about innovative ways to work at the nexus of AI and culture,” said Deborah Nelson, Dean of the Arts & Humanities. “UChicago is uniquely positioned to be the leading voice in national discussions of how emerging AI technologies can positively advance humanistic research, and, at the same time, how humanistic expertise in analyzing and understanding information in historical and cultural contexts can help catalyze the next generation of breakthroughs in AI.”

The project held its first workshop at the Neubauer Collegium Oct. 17-18. The event brought together nearly two dozen scholars representing 12 institutions, including UChicago faculty from five departments along with graduate students and postdoctoral colleagues.

“Everyone was able to get onto the same page and get to work,” said Long, the Andrew W. Mellon Professor of Japanese Literature and Digital Studies. “This was a rare opportunity for us to brainstorm research ideas that transcend any one discipline, and for those of us in the humanities and computer sciences to connect on an intellectual level. That made the whole event really exciting.”

Lightning Round

The workshop opened with a “lightning round” of brief talks in which each participant presented their current research and shared insights on how AI is affecting their respective fields. In the following breakout sessions, groups representing a mix of perspectives coalesced around topics of mutual interest and proposed ideas for pilot projects to pursue together.

Out of that iterative process came three research nodes that will organize work for the remainder of the Humanistic AI project, which runs through June 2027.

One cohort will investigate the potential of large language models to address humanistic questions through simulated scenarios. Another group is examining how AI tools might assist in the process of knowledge discovery and creative production. The third group will explore the historical antecedents, philosophical implications and creative potential of AI “slop,” a term used to describe poor-quality AI-generated content that circulates widely online and increasingly in academic and other work settings as emails, essays and research papers.

Kennedy, a formal linguist with specialized interest in the indeterminacy of meaning, sees strong potential in the research trajectories taking shape. A member of the “simulations” cohort, he is keen to see what cutting-edge modeling techniques can reveal about human culture, behavior and interaction.

“There’s clearly the seed of something that could be really new and different here,” said Kennedy, the William H. Colvin Professor of Linguistics. “But there is also a challenge: how will we know whether we are learning something interesting about humans, rather than learning something about language models?”

Sharing Perspectives

Participants in the Humanistic AI workshop shared their perspectives at a public roundtable discussion on Oct. 18, organized as part of the 2025 Arts & Humanities Day. The annual event, which showcases the richness and variety of scholarship and arts programming in the Division of the Arts & Humanities, partnered with Chicago Humanities for the first time this year.

The wide-ranging discussion addressed how the explosion of interest in generative AI is shifting the landscape for humanistic scholars. Speakers noted a flood of new studies and paper submissions as AI diffuses through their fields. They also considered the proliferation of data tracking human interactions with chatbots and other companions, corporate investment in research driving AI innovation, classroom experiments with AI-inflected pedagogy and more.

“There are things that AI doesn't do very well that are opportunities for people in the humanities,” said Ted Underwood, a professor of information sciences and English at the University of Illinois. “AI is really good at writing lots of kinds of texts. It summarizes things well. But it has not done a great job of putting novelists out of work.”

These so-called “failure modes” of the technology open new possibilities for researchers. Some collaborators may be motivated to design more effective models for narrative generation or, conversely, to develop ethical arguments for caution amid the rush to simulate creativity. Others will be inspired to examine the limits of AI storytelling, a way to deepen our understanding about aspects of the creative process that are essentially human.

“In studying the differences between two kinds of language production systems—humans and language models—we can learn something about both of them, and the differences then become the basis for new insights,” Kennedy said. “Humanists are particularly good at this kind of comparative work, and our hope is that extending it to a comparison of humans and AI will lead to unexpected and exciting discoveries.”

The Humanistic AI team will reconvene in June 2026 to collaborate on works in progress. At a final session, tentatively set for spring 2027, they will present the results of their work and chart next steps.

This article originally appeared in UChicago News.