The move marks a deepening split among leading AI developers over how aggressively lawmakers should police the fast-evolving technology.
The company formally endorsed the legislation on Tuesday, arguing it strikes a balance between fostering innovation and addressing risks such as algorithmic bias and job displacement. The bill would require companies to conduct impact assessments for high-risk AI systems but stops short of mandating third-party audits or outright bans on certain uses.
“This is a thoughtful, targeted approach that protects consumers without stifling the next generation of AI breakthroughs,” an OpenAI spokesperson said in a statement. The endorsement comes as Massachusetts lawmakers weigh competing proposals in one of the most closely watched state-level AI debates in the country.
Anthropic, by contrast, has backed a more aggressive bill that would impose mandatory testing requirements and create a dedicated state oversight board with enforcement powers. That proposal, which has drawn support from consumer advocacy groups, would also require companies to disclose training data and allow individuals to sue over AI-caused harms.
The divergence reflects a broader philosophical rift in the AI industry. OpenAI, which has pushed for federal preemption of state laws, argues that a patchwork of state regulations could hamper development. Anthropic has taken a more cautious stance, warning that unchecked deployment of powerful AI systems poses existential risks.
Massachusetts has emerged as a key battleground for AI regulation in part because of its concentration of tech firms and academic institutions. The state’s legislature is expected to hold hearings on both bills in the coming weeks, with a final vote possible before the end of the year.
Critics of the OpenAI-backed bill argue it lacks teeth, noting that its impact assessment requirements rely heavily on company self-reporting. “Without independent verification, these reports become little more than public relations exercises,” said one consumer advocate who spoke on condition of anonymity.
Supporters counter that a lighter touch is appropriate for a technology still in its infancy and warn that heavy-handed rules could drive AI companies to friendlier jurisdictions. The outcome in Massachusetts could influence regulatory approaches in other states and potentially shape the contours of any future federal legislation.
