Duke faculty, students, trainees and staff are engaging with artificial intelligence in a multitude of ways. A new initiative led by Provost Alec D. Gallimore seeks to reflect this diversity of approach, encourage engagement with tools and services, and contribute to a global conversation about the new horizons that AI can open up for us, and how to keep it from exacerbating some of our greatest challenges.
Early last year, Gallimore first envisioned bringing together Duke faculty and academic leadership who are leading work that develops, uses, studies, or otherwise engages with AI. To lead this new initiative, he tapped Yakut Gazi, vice provost for learning innovation and digital education; Michael Pencina, vice dean for data science and chief data scientist at the School of Medicine; and Joseph Salem, Rita DiGiallonardo Holloway University Librarian and vice provost for library affairs.
What they all thought might begin as a small conversation quickly became Duke's first AI Summit, in April of 2024. More than 230 faculty and administrators registered, with another 130 people on the waitlist. Every school and unit at Duke was represented. There was clearly a huge demand for dialogue around AI and its impacts, issues, and opportunities.
Feedback heard at the event and ensuing conversations with Duke's academic leaders and faculty groups in the following nine months coalesced into four pillars of AI at Duke:
Life with AI
This pillar focuses on understanding AI's impact on society, culture, and daily life, exploring key topics such as teaching and learning, the future of work, the meaning of knowledge and of humanity, AI's impact on democratic systems, human-AI partnerships, AI and social media and more.
Trustworthy and Responsible AI
This pillar emphasizes ethical, transparent and accountable AI systems. There is great enthusiasm around the potential of AI among technology developers, but this enthusiasm should be counterbalanced with scrutiny of AI's potential impact on society - and particularly in sensitive fields such as healthcare. Building on Duke's leadership in this area, the university is creating an interdisciplinary cluster of expertise that advances research and applications promoting privacy, data integrity, and more.
Sustainability in AI
Minimizing AI's negative environmental impact will require developing more sustainable AI tools, frameworks and cloud computing platforms. Quantum computing, for example, may present significant opportunities for investment, potentially enhancing environmentally responsible AI practices and infrastructure in support of the Duke Climate Commitment.
Advancing Discovery with AI
This area focuses on leveraging AI to accelerate scientific discoveries at Duke. It involves the development and application of AI algorithms and models to analyze large and complex datasets, uncover hidden patterns, and predict future trends in all scientific fields.
These pillars are built upon a strong commitment to a Duke-wide AI infrastructure and foundation. In spring 2025, a steering committee and four pillar advisory committees were formed to identify Duke's priorities around AI and propose recommendations to guide how the pillars are developed and applied across the university. The committees form a body of over 60 faculty and staff from all disciplines and departments, who will thoughtfully develop strategy and direction for the future of AI at Duke, and additional working groups are being formed to help drive Duke forward as a leader in this evolving field.
Providing leadership and direction to the initiative is a steering committee co-chaired by Gazi, Pencina and Salem. Four faculty chairs will spearhead the work of the pillar advisory groups:
- Life with AI: Chris Bail, professor of sociology and computer science;
- Trustworthy and Responsible AI: Nita Farahany, Robinson O. Everett Distinguished Professor of Law & Philosophy;
- Sustainability in AI: Ricardo Henao, associate professor in biostatistics & bioinformatics and electrical & computer engineering; and
- Advancing Discovery with AI: Victoria Szabo, research professor of art, art history & visual studies and director of the Information Science + Studies Certificate Program.
"Duke's teaching, research, and service should address the world's biggest challenges and help improve the human condition," said Gallimore. "AI is clearly one of the biggest opportunities and toughest challenges that we will face over the next few years. The recent advances we have made in AI represent perhaps the largest breakthrough humanity has had for decades, but we know that these advances can create new challenges or exacerbate pre-existing inequities.
"Our AI initiative is designed to be collaborative, interdisciplinary and synergistic, because we know that the thorniest challenges lie in those spaces between disciplines."
There is already lots of activity happening at Duke related to AI - and there is a new resource to learn more and get involved in conversations, projects, and events. To see what Duke schools and units are doing with and about AI, and to learn more about this developing initiative, you can visit Duke's new AI website: ai.duke.edu .
