Hochul expands New York AI safety testing center | Magica
Hochul expands New York AI research center into statewide safety testing effort
Editorial Team
••📖5 min read
New York will connect more than 80 researchers with Empire AI computing resources to evaluate advanced AI systems and advise state oversight, expanding a center first announced in January.
New York plans to bring together more than 80 researchers to independently evaluate advanced AI systems.
The Binghamton-based center already had a safety research mission and $55 million in announced seed commitments.
Its research will inform state oversight as AI developer reporting requirements begin in January 2027.
New York Governor Kathy Hochul on November 10 announced a statewide AI safety research network through the New York Center for AI Responsibility and Research at Binghamton University, part of the State University of New York (SUNY). Hochul first announced the center in January; the expansion adds interim scientific leadership and a role advising the state's new AI oversight office.
The consequential change is a route for outside technical assessments to inform government scrutiny of developers' safety claims. According to the governor's announcement, the center will coordinate with the state's new Office of Digital Innovation, Governance, Integrity, and Trust, known as DIGIT, to review published safety frameworks and recommend stronger requirements.
The new policy assignment makes the center more than a university research initiative. Its assessments could help officials judge developers' disclosures, although the announcement does not establish how much access researchers will have to models or how findings will be made public.
Jeannette Wing, an interim co-director of NYCAIRR, in Columbia’s 2021 appointment portrait. Photo: Jeffrey Schifman / Columbia University. Source: Columbia University.
A research network with a policy assignment
The center, abbreviated NYCAIRR, will bring together more than 80 researchers across New York, supported by Empire AI, the state's public-private computing initiative. Researchers will remain at their home institutions rather than all moving to Binghamton.
Interim co-director Greg Morrisett is dean and vice provost of Cornell Tech, Cornell University's research center. He previously led Cornell's Faculty of Computing and Information Science; his research on provably correct and secure software supplies relevant expertise for evaluating whether complex systems behave as intended.
Alibaba’s Qwen4-Max preview gives developers a fixed model to test through November 25. API access requires Token Plan, while claimed coding gains remain unquantified in the launch materials.
His fellow interim co-director, Jeannette Wing, is Columbia University's executive vice president for research. She previously led its Data Science Institute, which supported interdisciplinary research and access to real-world datasets, according to her 2021 appointment announcement. The governor's office says the two computer scientists were chosen to guide the scientific agenda, oversee independent evaluations and build the statewide network.
The remit extends beyond checking companies' published policies. Researchers are to develop evaluation tools and standards, conduct foundational AI science, use AI for scientific discovery and train AI safety experts. A national recruitment effort will seek expertise including mathematical foundations, formal verification, agentic systems, hardware security, privacy, interpretability, alignment and biosecurity.
These are research objectives, rather than reported testing results. The announcement does not specify a first set of models, an evaluation schedule or a measure of how many systems the center can assess.
January’s announced commitments paired a $30 million donor-group gift with $25 million in SUNY research capital, separately from Empire AI’s broader funding. Source: New York Governor's Office.
The money predates the statewide launch
The center's initial financing was announced on January 21: a $30 million gift from a donor group led by Tom Secunda, a Binghamton alumnus, Bloomberg LP co-founder and Empire AI board chair, paired with $25 million from SUNY. The original announcement identifies SUNY's contribution as a research capital investment. Those commitments total $55 million in seed funding. November's announcement also mentions state operating capital without specifying an amount or an annual operating budget.
Safety testing was part of the January vision: Secunda described independent research to test systems and measure performance, and the center's stated goal included foundational research into safety, security and transparency. November extends that mission through the statewide network and coordination with regulators.
A video description of the launch identifies Secunda as a major Democratic Party donor. His philanthropic and board roles connect the center's backing with its computing provider. The announcements do not detail donor control over research priorities or rules protecting publication of results; they also provide no evidence that he will direct evaluations.
Empire AI is a separate, $500 million public-private research computing initiative, according to the governor's November announcement. A report on the launch puts its backing at roughly $340 million in public funding and $160 million in private investment. These figures describe the broader consortium, rather than an additional allocation to NYCAIRR.
Binghamton is a founding consortium member, while Empire AI's computing center is housed at SUNY's University at Buffalo, according to the January announcement. That existing connection helps explain the choice of host and infrastructure. The November release gives no dedicated computing allocation or cost per evaluation, leaving the center's testing capacity unquantified.
The statewide expansion follows January’s center announcement and precedes the registration deadline for large frontier AI developers. Source: New York Governor's Office.
January brings the first oversight test
The launch comes ahead of new requirements under New York's RAISE Act, the state's frontier AI safety and transparency law. The governor's November announcement distinguishes obligations applying to all frontier AI developers operating in New York from additional duties for large developers.
According to that announcement, large frontier developers must register through the newly opened state portal by January 1, 2027 and begin publishing safety frameworks in January. All frontier developers must publish model transparency reports and report critical safety incidents; the state describes a 72-hour incident-reporting requirement.
DIGIT, within the state's Department of Financial Services, is responsible for implementing the law. NYCAIRR's stated role is to supply expertise, assessments and evidence at the state's request. Reviewing safety frameworks is a concrete starting point; independently testing the systems behind those documents requires operational details still absent from the announcement.
The next evidence to watch is therefore both regulatory and scientific: what developers disclose, what systems researchers can examine, and what evaluations the center releases. Those details will show how far the statewide network can turn its promised independence into scrutiny that public officials can use.
Railway estimates stolen customer keys generated about $740,000 in unauthorized AI spending after an October intrusion. Secrets from up to about 100,000 workspaces could have been decrypted, and the company says it will cover losses providers decline to refund.
Anthropic expects retail investors worldwide to receive up to 20% of its IPO, potentially $12.9 billion at the proposed upper price. Access through six brokerage channels depends on demand and broker rules, while a separate technology-company investment would help finance vast infrastructure commitments and founders and a special trust retain control.
Anthropic's seven founders plan equal IPO share sales worth about $250 million each at the proposed midpoint. Their sales make up 2.9% of the public offering, while the company seeks much larger financing under a structure that limits public investors' voting influence.
Nvidia’s independent directors retained Wachtell Lipton partner Joon H. Kim for an investigation into reports of a foreign intelligence effort targeting co-founder and CEO Jensen Huang. The board previously expressed confidence in Huang; the appointment provides no findings or timetable.
RRE Ventures led the funding for Graphon’s enterprise memory infrastructure. The startup is pitching connected, persistent context across different agents, but benchmark gains leave deployment costs and commercial scale unresolved.
Goldman Sachs, a lead bank on Anthropic’s IPO, reportedly forecasts $685 billion in annual revenue by 2030. The Claude developer has substantial cash reserves and rapidly growing sales, but rising compute costs, competitive pricing and delivery risks complicate the enterprise growth case.
Memories Upload pays contributors for accepted first-person video and sells annotated data to robotics developers. Its million-hour collection claim leaves open how much footage improves robot performance, while contributors surrender ownership without royalties.
OpenAI has acquired AI travel concierge Odessia and is bringing its founder, former Sonder chief Francis Davidson, into its travel effort. The standalone service has closed, leaving its future role alongside ChatGPT’s existing travel partners unclear.
Anthropic says expected advances from future models have triggered its research-acceleration threshold. An outside assessment finds faster progress but does not certify compliance with that threshold, while the voluntary policy leaves consequential decisions about safeguards and continued development with the company.
OpenAI’s free business-profile beta adds branded agents for eligible U.S. ecommerce and online service firms. The agents can also serve as paid-ad destinations, while OpenAI controls distribution and businesses bear responsibility for outputs.
Teleport has made Beams generally available with task-specific infrastructure permissions and a kill switch. Its access-security platform supplies an enterprise foothold, but established competitors and custom pricing for key controls complicate the buying decision.
Anthropic’s amended IPO filing raises its Microsoft spending commitment while disclosing an exit right from a $34.9 billion Lambda agreement. Google’s larger commitment was already disclosed earlier, and expensive short-term computing continues to squeeze margins.
The cash deal would bring Oura’s existing AI and physician care partner under its ownership. The $159 annual integration already connects members to doctors, but geographic limits, extra service costs and unproven clinical benefits qualify the expansion.
Reddit is testing AI-organized review pages for roughly 10,000 products. The pages link back to users’ conversations, but lack purchase verification—a safeguard Amazon says it uses for its own AI review highlights.
The renamed MegaETH Foundation is funding its own products after an earlier shift away from its accelerator. Its December inference market must show how it improves on existing services that already route requests by price and reliability.
Cambridge startup Tendrils Compute has raised a Benchmark-led seed round to build processors based on interaction nets. Its backers see an opening in branching workloads, including AI agents, but theoretical parallelism has yet to establish a commercial hardware advantage.
Google and Artificial Analysis launched a 200-task test of AI agents working with visually dense professional files. Google's study finds close leading scores and sharply different API bills, as its new Carbon model reaches customers alongside the Argon alternative.
Google is widening paid access to Gemini 4 Carbon for complex coding, business and scientific work. Gemini API token rates double after November 30, while benchmark results show that its cost advantage depends on the task and reasoning setting.
MiniMax has released downloadable weights for H3-Image, a model combining image generation and editing at native 16-megapixel resolution. Local deployment offers an alternative to its paid API, but hardware demands and a restrictive community license qualify the promise of open access.
