Last fall, the Data Science Institute received more than 17 proposals for faculty-designed, interdisciplinary event programs designed to spark intellectual exchange and drive new research directions in data science and artificial intelligence across Columbia.
Five projects involving over 30 faculty and researchers from more than 20 departments, schools and institutes across Columbia have been awarded funding and logistical support to organize events throughout the 2026 calendar year.
Centered on the themes Reprogramming Life and Thinking Machines, this series of symposia, workshops, conferences and research discussions will shape novel research directions and catalyze new partnerships. Below, learn more about the awarded projects and the faculty that will be leading these exciting new research efforts.
Want to get involved? Submit to the Thinking Machines RFP by March 22, and sign up for our newsletter to stay up-to-date on DSI and Columbia AI events and funding opportunities.
AI and machine learning are rapidly reshaping how we investigate, model, and manipulate living systems, from uncovering new biological principles to rethinking what it means to truly understand a cell or organism. Reprogramming Life explores how computational models are enabling new forms of discovery while also addressing critical questions about safety and accountability. It considers the emerging possibility of “reprogramming” biological and human systems and the ethical, legal, and governance challenges that accompany these advances. By connecting scientific innovation with questions of trust, patient rights, and public health, this series examines how AI can responsibly complement human intuition in research, medicine, and society.
This project brings together faculty James L. David and Nabila El-Bassel from the School of Social Work with biomedical informatics faculty Lena Mamykina and Orson Xu to explore how generative AI is opening new possibilities for personal health and mental wellness. At a time when a substantial proportion of Americans live in a mental health workforce shortage area, AI technologies may be a promising way to offer more people scalable, low-cost, and continuously accessible support—particularly for underserved or hard-to-reach communities.
“We have large segments of the population who don’t have access to mental healthcare,” said project lead and DSI member Lena Mamykina. While it’s possible that many of these people “could benefit from engaging with AI,” she said, “We know there are enormous risks.”
Some of those risks have been documented in research that highlights serious concerns about safety, accuracy, equity, and appropriateness, especially when AI tools are used in sensitive contexts such as mental health or substance use. Professional organizations and regulators are calling for stronger guardrails, clearer standards, and more meaningful inclusion of people with lived experience to ensure these systems do not reinforce stigma, exacerbate inequalities, or cause harm. This series aims to chart responsible paths forward for designing and deploying generative AI in personal health.
As AI increasingly impacts clinical care, clinicians must increasingly grapple how to respond when, for example, AI-generated predictions diverge from professional judgment. “AI is transforming healthcare, but raises profound ethical, legal and social quandaries that this project will help us to address,” said program lead and bioethics professor Robert Klitzman. Medical professionals must also increasingly assess how probabilistic system outputs should be documented and disclosed, as well as how they can safeguard patient rights in an environment of ubiquitous data collection.
Working alongside bioethics colleague Charles Binkley and Florence Hudson, director of the Northeast Big Data Innovation Hub, this project is bringing together a broad range of experts to examine how both clinical care and the growing use of biomedical modeling via “digital twins” can also introduce new challenges related to patient privacy, security, and data governance.
This program will examine how institutions can responsibly steward AI in science and health. Given that research oversight mechanisms like Institutional Review Boards were designed to address risks from specific kinds of research, they may not currently be well-positioned to assess the kind of group or community harms that AI systems can sometimes create.
“This initiative creates a rare commons across Columbia through which clinicians, engineers, social scientists, humanists, and community members can think together about what it means to integrate AI responsibly,” said Sandra Soo-Jin Lee, Professor of Medical Humanities and Ethics.
Together with colleagues Biomedical Informatics professor Noémie Elhadad and Chris Wiggins in Applied Physics and Applied Mathematics, this Frontiers series will organize symposia, workshops, and roundtables designed to generate new models for ethical stewardship of data and application of algorithmic systems in order to enhance accountability, and public trust.
Thinking Machines programs focus on foundational questions about how AI systems reason, generalize, and generate knowledge—and how those systems are reshaping research across disciplines. It spans research that uses generative models to probe social and psychological processes as well as work that examines the deep interplay between AI and mathematics, from hypothesis generation to the testing and evaluation of scientific claims.
Apply to the open Thinking Machines RFP through March 22, and sign up for our newsletter to stay up-to-date on DSI and Columbia AI events and funding opportunities.
Through training workshops, symposia, and research discussions, (Fun)damental AI & Math will explore the close, reciprocal relationship between AI and mathematics, and create structured opportunities for collaboration across domains, ranging from engineering and business to the social sciences and biomedical research. Unfolding in two phases, the first will examin how AI is being used in mathematics, starting with a hands-on workshop on large language models followed by symposia that that both a research view and a pedagogical perspective. The second phase will turn to the mathematical foundations of modern AI, exploring core theoretical challenges, new research questions, and culminate in a grant workshop aimed to forge collaborations and develop proposals.
“We aim to identify the foundational mathematical and AI problems emerging within [a diverse range of] fields and spark collaborations that push both AI and mathematics forward while charting new research directions,” said the Data Science Institute’s Kriste Krstovski. Alongside computer science and mathematics colleagues Andrew Blumberg, Ivan Corwin and DSI Associate Director for Research Daniel Hsu, Krstovski has assembled experts from a range of institutions to share insights and innovations that will help spark future research directions and collaborations.
In political science, economics, and psychology, large language models are being used to classify content, generate experimental stimuli and simulate responses to surveys. This program focuses on generative AI in political science and psychology research, exploring both novel applications and key methodological questions that arise as generative AI is increasingly used to classify content, generate experimental stimuli, simulate responses, and design adaptive surveys.
“While GenAI allows us to explore new research questions that would have been difficult to answer even a few years ago, it introduces a layer of complexity. The purpose of this [work] is to unpack that complexity using innovative statistical tools,” said program lead Yamil Velez, Assistant Professor of Political Science.
Algorithmic Revolution in Life Sciences and Healthcare Research (2026)
Toward Abstraction, Causality, and Common Sense (2026)