AI Future at Duke's Triangle Summit

Source: today.duke.edu

Published on June 16, 2025

The Triangle AI Summit recently took place to encourage more AI involvement and cultivate leadership to handle the technology's possibilities and risks. The summit brought together leaders from academia and industry to explore the future of AI. According to Jun Yang, Bishop-MacDermott Family Professor of Computer Science at Duke, AI is here to stay, and we must figure out how to use it as an extension of human intelligence.

Provost Alec Gallimore hosted the Triangle AI Summit at the Washington Duke Inn on May 30. Duke Learning Innovation and Lifetime Education (LILE), Duke Libraries, Duke Community Affairs, and Duke School of Nursing organized the summit to promote AI engagement and develop leadership skills to manage its potential and risks. Gallimore emphasized the importance of dialogue and collaboration and building connections within the Triangle community.

The summit highlighted Duke's leadership in AI. Faculty across all schools are involved in AI research. A pilot project with OpenAI provides undergraduate students with a prepaid license to ChatGPT-4o. Tracy Futhey, vice president and chief information officer of Duke’s Office of Information Technology, mentioned looking forward to future developments in AI.

AI's Impact on Society

Cade Metz, a New York Times technology reporter, was the keynote speaker. Panels featuring researchers, industry experts, and educators discussed AI's effects on society and its role in scientific innovation. The summit featured a student panel, a teaching showcase with over 20 AI demos for classroom use, and two interactive workshops on AI assistants. The sessions covered AI's societal benefits, potential dangers, and impact on the workforce and higher education.

AI in Healthcare

Nicoleta Economou-Zavlanos, director of Duke Health’s AI Evaluation and Governance program, shared examples of AI's impact on healthcare, noting AI's potential to detect strokes in brain scans more effectively than humans. She also mentioned the advantages of ambient technologies, which transcribe conversations between providers and patients, reducing note-taking and clinician burnout.

Brinnae Bent, executive in residence at the Pratt School of Engineering, discussed an AI-assisted medical device that aids individuals with neurological conditions like multiple sclerosis and cerebral palsy in regaining mobility. Bent also addressed AI's harmful impacts, such as deepfake pornography targeting teens and racially biased police technologies. Yakut Gazi, Duke’s vice provost for learning innovation and digital education, stated that AI disproportionately affects women in the workforce.

Humans and AI in the Workforce

Speakers agreed that AI is not yet poised to replace humans in most jobs. Bent emphasized the importance of humans in the workforce, referencing the fintech company Klarna's replacement of customer service representatives with AI, which led to a significant devaluation of the company. Metz urged the audience to consider AI’s flaws, noting that humans excel at handling chaos and unexpected situations, while machines are better at pattern recognition. He added that AI uses probability to generate results, which can be unreliable due to the vast amounts of data it pulls from the internet.

Chris Bail, professor of computer science and sociology and director of Duke’s society-centered AI initiative, warned that this can fuel the spread of misinformation. Yang stressed the importance of critical thinking when using generative AI. Joseph Salem, Duke’s vice provost for library affairs, highlighted the need to prepare students for a rapidly changing research environment. Student panelists from Duke’s Code+ program echoed the need for higher education to adapt. Duke sophomore Dara Ajiboy said that it’s about evolving and shifting the education system.

Duke’s AI steering and advisory committees will begin new work around the university’s AI Framework in the fall, creating new opportunities for the broader campus community to engage deeply with AI. To learn more, visit ai.duke.edu.