New York state lawmakers have moved one step closer to reigning in harmful artificial intelligence with the RAISE Act. This piece of legislation would require the makers of frontier AI models to meet transparency standards. The bill specifically targets major players in the AI industry, including leading labs in California and China, ensuring that these organizations adhere to stringent safety and security standards.
The RAISE Act was introduced with the help of New York state Senator Andrew Gounardes and New York Assemblymember Alex Bores. Bipartisan support Having passed through the New York state Senate with overwhelming bipartisan support, the bill now awaits consideration by Governor Kathy Hochul. She can either sign the budget into law, ask for changes in specific areas, or veto the budget in full. If enacted, the RAISE Act would mark a pivotal moment in the regulation of AI technologies in the United States.
The bill would also establish requirements for the world’s largest AI labs to release detailed safety and security evaluations of their most powerful frontier AI models. This entails heavy-hitters such as California’s OpenAI and Google, as well as Chinese counterparts DeepSeek and Alibaba. According to Anjney Midha, a general partner at Andreessen Horowitz, the RAISE Act aims to regulate these influential companies effectively.
“The window to put in place guardrails is rapidly shrinking given how fast this technology is evolving,” stated Senator Gounardes, emphasizing the urgent need for regulations in the rapidly advancing AI sector.
We call these provisions in the RAISE Act what they are: California’s controversial safety bill SB 1047 on steroids. New York Attorney General’s office tech companies up to $30 million in civil penalties. This is the power we should be using when those companies don’t live up to the expectations they’re held to. Second, despite rhetoric to the contrary, the legislation does not legally mandate AI model developers include a “kill switch” in their models. Moreover, those companies that model train frontier AI models thereafter will not be penalized for these critical harms.
Safety advocates from across the country are coming together to support the bill. Among those signers are Nobel Prize laureate Geoffrey Hinton and AI research pioneer Yoshua Bengio. We’re thrilled that they call the RAISE Act one of the principal victories for our AI safety movement. Yet this crucial movement has faced tremendous headwinds in recent years.
Even with the warm welcome from many in the development community, there are still critics of the bill. Midha expressed concerns about its potential impact, claiming, “The NY RAISE Act is yet another stupid, stupid state level AI bill that will only hurt the US at a time when our adversaries are racing ahead.”
Anthropic, another major player in the AI field, has not officially weighed in on the legislation. As far as the long-term impact of this development, co-founder Jack Clark said that the company is still assessing its implications.
Assemblymember Bores responded that although political dynamics may play a role in determining the bill’s fate, he is still hopeful about the bill’s prospects. “I don’t want to underestimate the political pettiness that might happen, but I am very confident that there are no economic reasons for them to not make their models available in New York,” he asserted.
