Massachusetts lawmakers are weighing some of the toughest state-level artificial intelligence rules in the U.S.
The state Senate passed an economic development bill last month requiring AI labs to go through independent reviews of the "catastrophic risks" every 120 days, Bloomberg reported. The results would be publicly released, but the state would not have the authority to use them to stop AI development.
The border legislation is expected to pass in the legislature prior to the November election, due to the fact that it includes funding for projects in lawmakers' home districts. A six-member negotiating committee made up of the House and Senate members is now working to resolve the remaining differences, including the AI provisions.
Bloomberg noted that Anthropic retained Boston-based lobbying firm Tremont Strategies Group earlier this year and OpenAI recently hired lobbying shop Benchmark Strategies to lead the effort.
Cesar Fernandez, Anthropic's head of U.S. State and Local Government Relations, stated in a post on X that the company is in favor of the new laws: "We applaud the Massachusetts Senate for passing the clearest and strongest AI safety legislation in the country. The Senate's economic development bill raises the bar on AI safety while ensuring that innovation is able to continue."
"We have long believed frontier AI safety is best addressed by a strong federal standard. But powerful AI won't wait for consensus in Washington and Massachusetts just showed what thoughtful state leadership looks like," he continued.
Meanwhile, OpenAI has cautioned the reviews could delay the rollout of cybersecurity-focused models that could address some of the threats lawmakers are concerned about. The ChatGPT maker is pushing for states to follow a consistent framework modeled after Illinois law.
"Different rules don't make things safer. They just create confusion," Donnie Fowler, OpenAI's head of U.S. state policy and partnerships, stated.
Illinois' artificial intelligence law calls for a third-party audit once a year, and is aimed at ensuring that AI companies are following the safety guidelines and disclosures outlined in the law. The Massachusetts proposal is similar, also requiring an annual compliance review; however, third-party groups would also have the authority to focus on risks tied to potential catastrophic events, including incidents that could kill or injure at least 50 people or cause more than $1 billion in property damage, Bloomberg noted.
Massachusetts State Senator Barry Finegold argued that independent testing is essential to making sure that frontier AI models don't pose serious risks to the public. If the state moves forward with the bill, he said other states are likely to also adopt this law.