Gov. Kathy Hochul of New York was expected to sign legislation on Friday that would establish new state new rules governing the safe development of the most-advanced artificial intelligence models.
The law, which state lawmakers originally approved in June, was the subject of intense lobbying before and after its passage by technology companies seeking to weaken its provisions.
The companies wanted New York’s A.I. regulations to more closely match those recently adopted by California, and successfully pushed Ms. Hochul to remove measures that would have subjected more companies to the new regulations and required them to take additional steps to ensure that the models they are developing are safe.
After intense negotiations, the governor agreed to sign the original bill and lawmakers agreed to approve the changes to the legislation that she has requested early next year after they return to Albany.
“New York is once again leading the nation in setting a strong and sensible standard for frontier AI safety, holding the biggest developers accountable for their safety and transparency protocols,” Ms. Hochul said in a statement adding that the reforms represent progress “as the federal government lags behind, failing to implement common-sense regulations that protect the public.”
The bill, the Responsible Artificial Intelligence Safety and Education, or Raise, Act, would be the first A.I.-related state law to be enacted since President Trump signed an executive order last week aimed at reining in states’ ability to limit the artificial intelligence industry. His order prompted bipartisan criticism and will most likely be challenged in court.
Mr. Trump has talked about creating national standards for certain A.I. models, but has not followed through. State legislatures have tried to fill the vacuum: Every state this year introduced A.I. legislation and nearly 40 adopted about 100 laws between them related to the technology, according to the National Conference of State Legislatures.
California and New York, two of the nation’s leading A.I. hubs, have moved aggressively to try to set up guardrails as the most advanced A.I. models continue to develop. New York’s efforts build on a California law that Gov. Gavin Newsom signed in September.
Companies like Meta, OpenAI and Google were able to shape the legislation in both states by persuading elected officials that burdensome regulations could hinder growth in a sector that is reshaping the national economy. They also were able argue convincingly that states should strive to align their laws so as not to create a regulatory patchwork. (The New York Times has sued OpenAI and Microsoft, claiming copyright infringement of news content related to A.I. systems. The companies have denied those claims.)
“With these new safeguards, New York is raising the bar,” said Sarah Heck, head of external affairs at Anthropic, the maker of the chatbot Claude and one of the leading companies behind the artificial intelligence boom. Last month, the company announced that Chinese state-sponsored hackers conducted a largely automated cyberattack using its A.I. technology.
“The fact that two of the largest states in the country have now enacted A.I. transparency legislation signals the critical importance of safety and should inspire Congress to build on them,” Ms. Heck said.
Chris Lehane, OpenAI’s chief global affairs officer, applauded Ms. Hochul’s move.
“While we continue to believe a single national safety standard for frontier A.I. models established by federal legislation remains the best way to protect people and support innovation, the combination of the Empire State with the Golden State is a big step in the right direction,” he added.
Some A.I. safety advocates and the New York bill’s sponsors had wanted, among other things, provisions that would prohibit companies from releasing models that do not pass the safety tests of the firms. They also wanted the bill’s rules to cover a far broader tranche of companies developing A.I. tools and to include much higher fines than California’s recently passed law.
Much of this was not in California’s bill and has now been removed from the New York one at the behest of Ms. Hochul and the technology industry. Still, the bill’s sponsors were pleased that they had augmented the California bill and hoped other states could build on New York’s work.
“Today is a major victory in what will soon be a national fight to harness the best of A.I.’s potential and protect Americans from the worst of its harms,” said Assemblyman Alex Bores, a Democrat who sponsored the bill with State Senator Andrew Gounardes, also a Democrat.
Mr. Gounardes said this legislation was the hardest bill to get passed and signed in his seven years in the State Legislature.
“Many in the tech community had mobilized significantly against this bill, and they were spending millions of dollars to lobby against it,” he said.
He added: “This shows that you can enact good policy, even if it’s incremental policy, that still is meaningful in the face of such concentrated power and opposition.”
The State Department of Financial Services will now create an office with rule-making authority that is focused on regulating A.I. and can require additional reporting from the companies. It will also issue reports on how the bill should be updated in the future. If companies fail to comply with these rules, the Attorney General will be able to bring civil actions against them with fines starting at $1 million.
The California law, known as S.B. 53, forces the most advanced A.I. companies to release safety protocols used in building their technologies and to disclose the greatest risks they posed. Mr. Bores explained that the Raise Act goes further by requiring more detail in “how they handle all of the risks.”
“That is a substantive difference, because now their safety plans need to be detailed enough that we know they are actually taking action,” said Mr. Bores, whose campaign for Congress faces financial opposition from these same technology companies.
The New York law also forces companies to report “critical safety incidents” far sooner than California will require.
Over the last two years, Ms. Hochul has invested more than $300 million in A.I. research and signed bills addressing artificial intelligence’s role in the film industry and elections as well as the mental health risks of chat bots.
She has met with a number of the sector’s leading figures and raised money from them for her re-election campaign. In October, she raised close to $250,000 at an event co-hosted by Mr. Bores, the bill’s sponsor, and a lobbyist working with A.I. safety proponents.
Ms. Hochul traveled to San Francisco and raised money for her campaign alongside Julie Samuels, the president of Tech:NYC, an industry association, and Ron Conway, a prominent Silicon Valley venture capitalist. Ms. Samuels and Mr. Conway, who both had raised money for Mr. Hochul in the past, had expressed concerns about the RAISE Act’s original language.
Ms. Samuels praised the revised bill in New York, calling it “responsible legislating.”
“Instead of trying two different things and creating competing regulatory frameworks, New York and California have shown real leadership by ensuring that our country is a place where we can balance innovation and safety.”
Benjamin Oreskes is a reporter covering New York State politics and government for The Times.
The post Hochul Reaches Deal on A.I. Regulation in New York appeared first on New York Times.
