New York Launches Public Center to Test AI Models and Opens Registration for Major Developers Under the RAISE Act
On November 10, Governor Kathy Hochul launched the New York Center for AI Responsibility and Research at Binghamton University, bringing together over 80 researchers to independently evaluate advanced AI models. The state also opened the portal where major frontier model developers must register before January 1, 2027.
“We cannot just take AI companies at their word.” With this remark, New York Governor Kathy Hochul launched the New York Center for AI Responsibility and Research (NYCAIRR) on November 10, a public center tasked with independently testing the most advanced AI models. On the same day, the state opened the portal where major developers must register under its RAISE Act.
A Statewide Evaluation Lab
Based at Binghamton University, part of the public SUNY system, the center connects over 80 researchers who remain affiliated with their home institutions and leverages the supercomputers of Empire AI, the $500 million public-private initiative launched by the state. It is co-directed on an interim basis by Greg Morrisett, dean of Cornell Tech, and Jeannette Wing, executive vice president for research at Columbia.
Three main missions were announced: testing the safety and security of frontier systems, developing the tools and standards needed to evaluate increasingly capable models, and training a new generation of specialists. The center is prioritizing recruitment in formal methods and verification, agentic systems, hardware security, interpretability and alignment, and biosecurity. First announced in January, it is funded by a $30 million donation led by Tom Secunda, co-founder of Bloomberg LP, along with $25 million from SUNY.
The RAISE Act Takes Effect
The center also plays a regulatory role. Upon state request, it can provide evaluations and expertise, and it will work with the DIGIT office—responsible for enforcing the RAISE Act—to review safety frameworks published by developers and recommend improvements. DIGIT opened the registration portal: major frontier model developers have until January 1, 2027, to sign up. Starting in January, they will be required to publish their safety frameworks, and all frontier model developers must report any critical incidents within 72 hours and publish transparency reports.
Our take
New York, which already regulated companion chatbots for minors, is equipping itself with a tool few states possess: public testing capacity backed by public compute. In the absence of federal legislation—a gap the governor emphasized—this creates a state-level model akin to national safety institutes. The known limitation remains: an evaluator is only as good as its access to models, and the announcement says nothing about pre-release access. The timeline, however, is concrete: major labs have less than two months to register, and the first safety frameworks published in January will give the center its initial testing ground.
Sources
A primary source comes directly from the company, lab or organization concerned. A secondary source reports the information without being its author: we flag it as such. Spotted an error? See our corrections policy.
