Students Explore AI’s Possibilities Through Interdisciplinary Research

Through Bass Connections, undergraduates, graduate and professional students, postdoctoral fellows and faculty are coming together to examine the role of AI across disciplines, discovering how it can bring distant campuses closer together, make academic publishing fairer, help us understand the past and offer insights into medical research and healthcare policy.

This year, nine project teams are participating in research that directly engages with AI. These interdisciplinary projects range from exploring how to protect the privacy of people living with dementia to considering what AI’s storytelling limitations tell us about how human creativity works.

Words & Ideas

Generative AI and Cultural Creativity

Generative AI can create stories, but can it create good stories? Scholars across literature, cognitive science and anthropology often argue that compelling storytelling depends on surprise, and generative AI still struggles to create narratives that don’t feel formulaic. This team is examining factors such as predictability, originality, coherence, pacing and engagement in stories to ask the question: What does AI’s difficulty creating plot twists and exciting narrative surprises reveal about the nature of human creativity?

Open Access Academic Publishing Using Game Theory and Graph Theory

This project is investigating whether new systems of scholarly credit and incentives can make academic publishing fairer and more transparent. Building on the open-source platform Liberata, the team will replace traditional ordinal authorship with contribution shares that reflect each researcher’s proportional contribution. Team members will use game theory and graph theory to analyze contributions and citation networks, and build responsible AI tools to support collaboration, manuscript discovery and reviewer matching.

Image

Advancing Innovation

Developing AI Applications for Traumatic Brain Injury Care

Each year, an estimated 5 million Americans seek emergency medical care for traumatic brain injury (TBI), but it can be challenging for clinicians to reliably predict outcomes for these patients. This team is developing, refining and evaluating an AI-driven clinical decision support tool that will help healthcare providers make better informed treatment decisions for TBI. By interviewing emergency physicians, neurologists, neurosurgeons and rehabilitation specialists, team members are trying to understand clinical workflows and attitudes toward AI-assisted care. Designers, product managers and engineering students will translate these insights and model outputs into a prototype clinical decision support tool.

Advancing Research Translation with AI Stakeholder Insights

Duke produces an extraordinary volume of scientific innovation, yet much of this work never reaches the public in the form of usable products, services or technologies. This project is building on the interdisciplinary strengths of Bass Connections to bridge the gap between research creation and real-world application. By pairing students with Ph.D. researchers and using AI tools for structured brainstorming, the team will help identify promising use-cases for emerging technologies and gather stakeholder insights to guide translation pathways.

Protecting Health & Privacy

AI Global Health: Policy, Evidence and Financing for Responsible AI

AI tools are increasingly used in clinics and public health programs across low-and middle-income countries (LMICs) because there are not enough staff or financial resources to meet medical needs. Algorithms support triage, guide treatment, predict outbreaks and help health systems allocate scarce resources. Yet this rapid adoption is outpacing the protections that typically ensure safety, accountability and equitable access. This team is working with international partners to close these gaps by developing a Responsible AI in Health Policy Toolkit and other practical resources to guide decisions about best practices on adopting, improving and scaling AI-enabled health technologies in LMICs.

Protecting Privacy of Persons with Dementia

AI companion technologies, like chatbots and smart speakers, are increasingly being used as supportive tools for dementia care. These systems offer potential benefits, including emotional support, reminders and social engagement. However, because commercial AI companions operate with little regulation or oversight, they could pose privacy risks. Through interviews with care partners, individuals with early-stage dementia and privacy experts, this team is investigating users’ privacy needs, expectations and concerns. Findings will inform the development of user-centered privacy protections, educational materials and public-engagement initiatives to help protect this vulnerable population.

Bringing Us Together

Building an Open Access Platform for Chinese Law and Policy

This project addresses the challenges that scholars of Chinese law and policy face in accessing, interpreting and sharing research materials. The team is investigating how an open-access platform can improve transparency, usability and cross-border collaboration. They’re building a searchable portal that curates and annotates scholarship, uses human-verified AI tools to organize metadata and visualizes transparency across studies. By comparing developments in copyright, licensing, privacy, open-access and AI governance in China, the U.S., Canada, the United Kingdom and the European Union, the team is advancing conversations about open scholarship.

The Duke-DKU Presence Lab Portal Project

As digital and physical environments increasingly blend, universities face a pressing question: how can communities experience meaningful presence across distance? This project engages with the emerging Duke-DKU Presence Lab to bring the two campuses closer together through technologies like virtual reality, telepresence systems, projection mapping, environmental sensing and AI. The team is designing, building and testing experimental “portals” — creative, technologically mediated experiences — that support synchronous and asynchronous exchange. The project will culminate in a joint Duke-DKU exhibition featuring co-located events staged simultaneously in both locations and mirrored in virtual space.

Connecting With the Past

Making Meaning at Historic Places

How can digital methods recover environments and histories that no longer exist? This team is investigating how historical evidence, experiences of place, and social and environmental justice shape the interpretation of historic landscapes and sites. Building on prior research at Catawba Trail Farm and Bennett Place in Durham, N.C., this year’s team is examining additional sites, including Duke Homestead State Historic Site, with a particular focus on labor and community connections. Using archival research, oral histories, AI-assisted analysis and digital storytelling, the project will develop new ways for communities to make meaning at historic places.

Image

Learn More

• How Does AI Map Meaning Across Languages and Time?
• Explore the complete list of 2026-2027 Bass Connections teams.
• Find out more about AI at Duke.

Main image: Clasping hands graphic courtesy of the Duke/DKU Presence Lab Portal Project.