The New York Center for AI Responsibility and Research at Binghamton University will conduct foundational research into AI safety, security, transparency, and responsible deployment. Gov. Kathy Hochul first announced the center during her State of the State address in January, with the formal launch coming now.
Who Is Running It
Leadership comes from two of New York’s leading research institutions, with more than 80 scientists expected to join the effort.
Jeannette Wing, Columbia University’s executive vice president for research, and Greg Morrisett, Cornell Tech’s vice provost, will lead the center. More than 80 scientists and AI-systems researchers from across the state are expected to contribute.
The Money Behind It
A philanthropic gift led by a Bloomberg LP co-founder and a separate state investment together represent the largest funding commitment in Binghamton University’s history.
Tom Secunda, a Bloomberg LP co-founder and Binghamton alumnus, led a donor group that contributed a $30 million philanthropic gift to anchor the center, which the university describes as the largest academic gift in its history. A separate $25 million research-capital investment from SUNY and New York State brings the total commitment to $55 million.
Where the Computing Power Comes From
Empire AI, the state’s existing public-private research consortium, will supply the computing infrastructure the center needs.
The center will draw computing capacity from Empire AI, New York’s roughly $500 million public-private AI research partnership. Large-scale systems housed at the State University of New York at Buffalo will support that work.
What the Center Is Meant to Do
The core mission is to give policymakers a source of independent technical evidence beyond what companies themselves provide.
Officials evaluating AI systems have typically depended on disclosures that companies prepare and control. The center is intended to supply independent assessments that could inform regulatory reviews, public procurement, and safety determinations, though those specific uses represent forward-looking goals rather than formally assigned functions.
The Access Problem Built Into the Design
Researchers face a structural limit: access to internal models and unreleased systems depends on what companies share voluntarily or what regulators can compel.
Unless firms volunteer access or regulators compel disclosure, internal models, development-stage systems, and unreleased capabilities remain off-limits to outside researchers. That constraint is a common limitation of AI oversight as it currently operates, not a flaw specific to this program.
A January Deadline Gives the Work Immediate Purpose
New York’s RAISE Act, taking effect January 1, 2027, creates a set of filings from large frontier AI developers that the center could examine.
The RAISE Act requires large frontier AI developers to establish and publish safety and transparency frameworks, report critical safety incidents within 72 hours, and register with the state. The center is expected to begin work before those requirements take effect, though the available sources do not confirm that evaluating RAISE Act filings is a formally assigned function.
The Larger Regulatory Vacuum
New York’s move comes as states fill a policy gap that federal authorities have not yet closed.
New York and California have both emerged as important regulatory centers amid unsettled federal AI policy. POLITICO also reports that congressional Democrats are preparing broader federal measures, including possible subpoenas of major AI executives and legislation covering online child safety and export controls.
Testing Is Not Enforcement
Independent research generates evidence; turning that evidence into consequences requires decisions that policymakers have not yet made.
Whether the center’s findings affect licensing, procurement, enforcement actions, or disclosure obligations depends on choices that New York officials have not yet announced. The center has a research mission, not automatic enforcement authority, and that distinction will shape how much practical weight its conclusions carry.
