Harvard Kempner Institute

The Kempner Institute for the Study of Natural and Artificial Intelligence at Harvard University is Harvard's flagship AI research institute, founded in December 2021 by a $500 million donation from Priscilla Chan and Mark Zuckerberg through the Chan Zuckerberg Initiative. The institute is named for Yael Kempner, Priscilla Chan's mother, and operates research at the intersection of natural intelligence (neuroscience and cognitive science) and artificial intelligence (machine learning and AI), with substantive cooperation across Harvard's medical school, neuroscience departments, computer science, statistics, and adjacent academic units. The institute is co-led by Bernardo Sabatini (Harvard Medical School neuroscience professor) and Sham Kakade (Harvard computer science and statistics professor; former Microsoft Research and University of Washington researcher). As of April 2026, the Kempner Institute is one of the principal academic AI research institutes globally, with the substantive parent-university funding scale and the natural-and-artificial-intelligence research focus that distinguishes the institute from peer academic AI research organizations.

At a glance

• Founded: December 2021 at Harvard University by $500 million donation from Priscilla Chan and Mark Zuckerberg through the Chan Zuckerberg Initiative.
• Status: Research institute within Harvard University.
• Funding: $500 million founding donation. Subsequent operational funding from Harvard, research grants, and adjacent academic-research funding.
• CEO / Lead: Co-led by Bernardo Sabatini, Harvard Medical School neuroscience professor, and Sham Kakade, Harvard computer science and statistics professor.
• Other notable leadership: Bernardo Sabatini, Co-Director. Sham Kakade, Co-Director. Senior research-leadership across the institute's research-program areas.
• Open weights: Yes, partial. Selected research outputs released open-source through GitHub.
• Flagship outputs: Active publication record at NeurIPS, ICML, ICLR, neural-systems-and-cognition venues, and adjacent computational-neuroscience and AI research venues; substantive cooperation across Harvard's medical school, neuroscience departments, and computer science programs.

Origins

The Kempner Institute was established in December 2021 with the founding announcement of a $500 million donation from Priscilla Chan and Mark Zuckerberg through the Chan Zuckerberg Initiative. The donation was characterized at the time as one of the principal individual philanthropic AI research-funding contributions globally, alongside the Schmidt Sciences parent-foundation contributions and other comparable philanthropic AI research-funding programs.

The institute's research-program mandate was explicitly oriented around the intersection of natural intelligence (neuroscience, cognitive science) and artificial intelligence (machine learning, deep learning), with the founding research thesis that the integrated study of biological and artificial intelligence systems would produce research insights that single-discipline approaches cannot match. The naming for Yael Kempner (Priscilla Chan's mother) anchored the substantive personal-foundation positioning of the donation.

The 2022 leadership recruitment included Bernardo Sabatini (Harvard Medical School neuroscience professor with substantive research output on synaptic plasticity and dopamine systems) and Sham Kakade (Harvard computer science and statistics professor; previously a senior researcher at Microsoft Research and the University of Washington with substantive ML-theory research output) as Co-Directors. The Co-Director structure reflected the natural-and-artificial-intelligence integrated research mandate.

The 2022 to 2026 period has built substantive research-program infrastructure across Harvard's research-collaboration network, with research output at NeurIPS, ICML, ICLR, neural-systems-and-cognition venues, and adjacent computational-neuroscience and AI research venues. The institute's research-cooperation relationships across Harvard's medical school, neuroscience departments, and computer science programs have been characterized in industry coverage as one of the principal Harvard cross-disciplinary research-cooperation programs.

Mission and strategy

The Kempner Institute's stated mission is to understand the basis of intelligence in natural and artificial systems, with the integrated research approach combining neuroscience, cognitive science, machine learning, and applied AI. The strategy combines two threads. First, foundational research on natural and artificial intelligence with focus on the principles of learning, representation, and computation that span biological and artificial systems. Second, substantive cooperation across Harvard's medical school, neuroscience departments, computer science programs, and adjacent academic units that anchors the cross-disciplinary research-collaboration positioning.

The competitive premise of the natural-and-artificial-intelligence integrated approach is that research insights from biological intelligence systems can advance artificial intelligence research and vice versa, with the principal academic-research advantage being the substantive Harvard research-collaboration network and the multi-decade neuroscience research lineage that the institute draws on.

Models and products

• Active academic-publication program. Research publications at NeurIPS, ICML, ICLR, neural-systems-and-cognition venues, computational-neuroscience venues, and adjacent AI research venues.
• Cross-disciplinary research-cooperation across Harvard. Research-collaboration relationships with Harvard Medical School, the Department of Molecular and Cellular Biology, the Department of Computer Science, the Department of Statistics, the Center for Brain Science, and adjacent Harvard academic units.
• Research-talent pipeline. Postdoctoral research-fellow programs and adjacent research-talent-pipeline programs.
• Selected open-source contributions. Through the Kempner Institute GitHub organization.

Distribution channels include academic-publication, cross-disciplinary research-cooperation, and the academic-research-talent pipeline.

Benchmarks and standing

The Kempner Institute's evaluation framework focuses on academic publication metrics, the substantive cross-disciplinary research-cooperation impact, and the broader research-talent pipeline that the institute supports.

Industry coverage has consistently characterized the Kempner Institute as one of the principal philanthropically funded academic AI research institutes globally, with the $500 million founding donation, the natural-and-artificial-intelligence integrated research mandate, and the substantive Harvard cross-disciplinary research-cooperation network as principal validating data points.

Leadership

As of April 2026, the Kempner Institute's leadership includes:

• Bernardo Sabatini, Co-Director. Harvard Medical School neuroscience professor.
• Sham Kakade, Co-Director. Harvard computer science and statistics professor.
• Senior research-leadership across the natural and artificial intelligence research-program areas.

Funding and backers

$500 million founding donation from Priscilla Chan and Mark Zuckerberg through the Chan Zuckerberg Initiative. Subsequent operational funding from Harvard University, external research grants, and adjacent academic-research funding.

Industry position

The Kempner Institute occupies a distinctive position as one of the principal philanthropically funded academic AI research institutes globally, with the substantive founding-donation scale, the natural-and-artificial-intelligence integrated research mandate, and the substantive Harvard cross-disciplinary research-cooperation network. Industry coverage has consistently characterized the institute as one of the structurally consequential academic AI research organizations in the broader US academic AI ecosystem.

Competitive landscape

• MIT CSAIL, Stanford AI Lab, Stanford HAI / CRFM, Berkeley BAIR, CMU SCS. US academic AI research peers.
• Princeton Language and Intelligence, NYU CDS, USC ISI, UW Allen School / NLP. Adjacent US academic AI research peers.
• Allen Institute for AI (Ai2). Adjacent philanthropically funded AI research institute.
• Schmidt Sciences, Open Philanthropy. Adjacent philanthropic AI research-funding peer organizations.
• Harvard Medical School, Center for Brain Science. Cross-disciplinary research-cooperation partners within Harvard.

Outlook

• The continued cross-disciplinary research-cooperation across Harvard medical school, neuroscience, and computer science programs through 2026 to 2027.
• The continued academic-publication output at NeurIPS, ICML, ICLR, neural-systems-and-cognition venues, and computational-neuroscience venues.
• The continued postdoctoral research-fellow program and broader research-talent pipeline.
• The continued natural-and-artificial-intelligence integrated research direction.

Sources

• Harvard Kempner Institute. Institute reference.
• Bernardo Sabatini Wikipedia. Co-Director reference.
• Chan Zuckerberg Initiative. Founding-donation source.
• Harvard University. Parent university reference.