New York launches independent AI testing, with access limited to public disclosures

New York is giving researchers a state-backed role in checking powerful AI systems rather than relying on developers’ safety assurances. Governor Kathy Hochul launched the program on November 10, 2026, through a new center at Binghamton University. Its assignment is to test advanced systems, investigate risks and give public officials evidence for their decisions.

The team behind the tests

The New York Center for AI Responsibility and Research will bring together more than 80 New York scientists. Hochul’s office describes the initiative as the nation’s first state-run AI testing program, positioning independent researchers as a check on the companies developing the technology.

Interim co-directors Greg Morrisett of Cornell Tech and Jeannette Wing of Columbia University will lead the center. Its stated remit includes testing advanced AI systems for both safety and security, alongside research intended to improve understanding of AI itself.

Researchers will draw computing power from Empire AI, New York’s $500 million public-private AI partnership. That figure describes the partnership supplying computing resources, not a separately announced budget for the new research center.

The mission extends beyond finding problems. Hochul’s office says the center will also work on using AI responsibly in medicine and scientific discovery, and on accelerating discoveries across scientific fields. Those are research goals for the center, rather than findings or benefits demonstrated by the launch.

The stakes are too high to simply take AI companies at their word.

Governor Kathy Hochul, in a press release

Independent scrutiny has an access limit

The program can study only information AI companies disclose publicly, according to Engadget. Independent researchers and computing resources therefore do not, by themselves, give the center access to unreleased models. That is a material boundary on what this new layer of scrutiny can examine.

New York’s existing RAISE Act will require more disclosures. Large AI developers must register with the state by January 1, 2027, then publish safety frameworks and transparency reports. They must also report critical safety incidents within 72 hours, adding an incident-reporting obligation alongside the public documents.

Those requirements provide regulatory context for the center’s work, but they do not close the access gap for unreleased systems. The launch establishes an independent testing and research program; it does not give researchers unrestricted access to everything an AI developer is building.