Kathy Hochul proposes major changes to New York's AI bill amid intense tech lobbying pressure

Kathy Hochul Redlines RAISE Act With Major Revisions

New York Governor Kathy Hochul has delivered a significant blow to the state's AI safety bill this week by proposing sweeping changes that would dramatically weaken its oversight provisions. According to sources who reviewed the redlined document, the original RAISE Act has been crossed out with replacement text that nearly verbatim resembles SB53, a California AI bill that Governor Gavin Newsom signed earlier this year 1

1

. The Responsible Artificial Intelligence Safety and Education Act, initially authored by state Senator Andrew Gounardes and Assemblymember Alex Bores, was written to require incident reporting and safety plans for powerful AI models developed in New York 1

1

.

The state's AI safety bill no longer includes several critical components that made it one of the more stringent proposals for state-level AI regulations in the country. Per a person familiar with the negotiations, the revised version has removed a requirement for detailed AI safety standards from AI companies that would reasonably reduce the risk of harm, a ban on releasing AI models that pose an unreasonable risk of harm, and requirements to report security incidents 1

1

. The bill also features a weakened catastrophic risk disclosure component, and provisions that would have applied to future models of foreign AI systems like DeepSeek have been eliminated .

Tech Lobbying Intensifies Around New York Legislation

The dramatic shift in the AI bill comes amid intense tech lobbying pressure on state lawmakers. This week, pro-AI PAC Leading the Future launched advertisements targeting Bores in his district NY-12, claiming the RAISE Act would "fail to keep people safe" and "crush jobs" 1

1

. Bores, who is running for former Representative Jerry Nadler's Congressional seat, has been pushing for Hochul to sign the original bill, making him a focal point in the national debate over how states should regulate artificial intelligence 1

1

. AI companies and startups have been urging major changes to the legislation, specifically suggesting it more closely mirror the California AI bill that Hochul now appears ready to adopt 1

1

.

The California model that Hochul is looking to replicate requires large AI developers to make public disclosures about safety protocols and report safety incidents, while also creating whistleblower protections and expanding cloud computing access for smaller developers and researchers 1

1

. This lighter-touch approach has drawn support from the tech industry but criticism from those who believe stronger safeguards are necessary as AI models grow more powerful.

National AI Regulation Debate Intensifies

The controversy surrounding the RAISE Act reflects broader tensions about whether states or Congress should set the rules for AI development. Hochul herself previously flagged concerns about a patchwork of state regulations, stating on Bloomberg TV months ago that "people prefer to have a federal regulation" because "it's hard when one state has a set of rules, another state does, another state. I don't think that's a model for inspiring innovation" 2

2

. Critics argue that if AI developers must cope with 50 different sets of contradictory regulations, America risks losing the race with China to set global standards for artificial intelligence, potentially paying a huge long-term economic price 2

2

.

Opponents of the original RAISE Act have raised particular concerns about provisions they say would boost trial lawyers at the expense of American innovation. Critics point to language that could allow lawsuits against one tech company over how another programmer uses open-source code, effectively discouraging US companies from making their designs open-source even as this practice helps American software become the global standard 2

2

. Some observers warn that lawsuit-friendly laws combined with New York's high taxes and energy costs could push cutting-edge tech companies to states like Texas and Florida 2

2

.

What Comes Next for State-Level AI Regulations

Hochul has ten days, not counting Sunday, from the time she receives the bill to sign, veto, or agree to "chapter amendments" with bill sponsors 1

1

. The decision carries significant implications beyond New York, as states increasingly become venues for intense debates over AI governance while President Trump pushes ahead with plans to stop state-level regulation through an executive order 1

1

. The tension between protecting intellectual property rights—including copyright protections so AI companies cannot mine publishers' content without compensation—and avoiding regulations that strangle American tech companies remains unresolved 2

2

. As the debate continues, observers are watching whether Congress will step in with national AI regulation that balances consumer protection with the need to maintain American competitiveness in the global AI race.