The inaugural CoreAI Discovery Day featured research, discussions and demonstrations related to artificial intelligence from faculty members and students in the department of computer science.
Robots that autonomously navigate needles through living tissue. AI systems that understand video better than proprietary models like GPT-4o. Deep brain implants that adapt in real time to treat Parkinson’s disease. These are not distant possibilities — they are active research projects in UNC-Chapel Hill’s Department of Computer Science, and on February 9, the department opened its doors to show the campus what foundational AI research looks like and what it can enable.
CoreAI Discovery Day, the department’s inaugural AI research showcase, filled Sitterson Hall with researchers from across the university — faculty members from chemistry, geography, philosophy, pharmacy, psychiatry and radiology alongside computer scientists, graduate students, UNC alumni, and university leadership. Chancellor Lee Roberts delivered opening remarks and toured five spotlight research demonstrations before the formal program began.
“A key priority for me as chancellor is positioning Carolina to lead in this new era of artificial intelligence,” Roberts said. “It’s encouraging to see the computer science department already at the forefront, building strong interdisciplinary partnerships across campus. Events like CoreAI create the space for that kind of collaboration.”
The event was conceived to answer a question that Department Chair James Anderson posed in his opening welcome: while headlines focus on using AI tools like ChatGPT, who is building the foundations that makes those tools possible?
“AI, machine learning, and related research areas have been prioritized by this department since before many of the modern public-facing AI tools existed,” Anderson said. “We see the applications of AI all around us, but many of those applications trace their origins to work done by our faculty and students going back years and even decades.”
From algorithms to action
The day featured twelve “lightning” talks across three sessions, organized around the areas where UNC computer scientists are advancing the field: computer vision and video intelligence, machine learning foundations and language intelligence, and robotics and embodied intelligence.
Distinguished Professor Mohit Bansal, who co-hosted the event and provided an opening primer on the department’s AI research portfolio, framed the scope of the work underway. The department, ranked eighth nationally since 2024 in natural language processing, machine learning and computer vision by CSRankings, has grown its AI-focused faculty significantly in recent years, adding researchers in areas from biometrics and deepfake detection to cyber-physical systems and society-centered AI.
“UNC CS undertakes research at every level of the AI stack, from the mathematical foundations up to the user-facing applications,” Bansal said. “AI is a university-wide priority, and it has been a pillar of our department’s research program for many years now. Our hope is that events like this one can be a catalyst for collaborations that extend our core AI work into projects across campus.”
The talks made the research tangible. Assistant Professor Gedas Bertasius described how his lab’s video understanding framework — which outperforms commercial models from Google and OpenAI — could give everyone access to world-class coaching by providing personalized feedback on skills from basketball to piano. Distinguished Professor Xiaoming Liu, who joined UNC in January, presented his work helping humans distinguish genuine media from AI-generated fakes, an increasingly urgent challenge in the era of publicly available generative AI.
In the robotics session, Distinguished Professor Ron Alterovitz showed how his Computational Robotics Group demonstrated the first AI-driven medical robot capable of autonomously steering a flexible needle through living tissue — outperforming human physicians in bronchoscopy procedures. Distinguished Professor Samarjit Chakraborty, working jointly with UNC Neurosurgery, presented AI-powered deep brain implants for personalized treatment of Parkinson’s disease that learn and adapt in real time.
Other talks explored building AI that millions of people can shape and govern, developing theoretical guarantees for when and why AI systems work, creating socially-aware language models for education and mental health, and deploying AI agents for precision agriculture.
Where disciplines connect
Five spotlight demonstrations in Sitterson Hall’s upper lobby gave attendees hands-on exposure to the research before and during the event. Visitors could see an autonomous surgical robot in action, watch AI reconstruct 3D models from endoscopy video, and interact with systems that reason about what’s happening in video more effectively than leading commercial products.
The afternoon proved equally significant. The Networking Lunch and AI Research Fair drew poster presentations from across the university — researchers from RENCI, the Eshelman School of Pharmacy, and the departments of chemistry, geography and the environment, statistics and operations research, and applied physical sciences, among others. Projects ranged from AI-designed protein therapeutics to machine learning for forest biomass mapping to adversarial attacks on financial forecasting models.
The cross-disciplinary energy carried into the closing panel, “Core to Application: How AI Research Enables Discovery Across Carolina,” moderated by Jay Aikat, vice dean of the School of Data Science and Society. Panelists from computer science, philosophy, data science and radiology explored what it actually takes to move from a foundational algorithm to a tool that changes real-world outcomes — and the critical role of thoughtful guardrails when AI systems affect real lives.
“The conversations that happened in the hallways were just as important as what happened on stage,” shared Senior Associate Dean for Natural Sciences Jaye Cable, who sponsored the event. “Researchers who had never met discovered overlapping interests. Faculty from other departments learned about computational methods that could accelerate their own work. That’s exactly what we hoped would happen.”
Building on momentum
CoreAI Discovery Day also marked a step in Chancellor Roberts’ vision for a coordinated, university-wide AI strategy. Vice Provost for AI and Chief AI Officer Jeffrey Bardzell, who delivered remarks at the beginning of the event, has been working to connect the research happening in computer science with applications across every school and college at Carolina.
The department plans to build on the connections made at CoreAI through ongoing partnerships and events.
“From its founding, this department has always been focused on using computer science to enable and enhance work in other fields across this campus and the state,” Anderson said. “I’m excited to see the new types of problems we can solve across campus after the new collaborations that will stem from this event.”
CoreAI Discovery Day was hosted by the UNC Department of Computer Science in the College of Arts and Sciences. For more information about the department’s AI research, visit unc.edu/unc-ai.
The department will hold a community open house on March 7, opening its doors more broadly and showcasing research across all of computer science — not just AI — to the wider Chapel Hill community. The CS Community Open House takes place Saturday, March 7, 2026.
By the Department of Computer Science