AI Frontier Institute

Promoting economic development with emerging artificial intelligence models and safety

View primary source →

Senate No. 2630 (reported favorably Oct. 16, 2025 by the Advanced Information Technology, Internet and Cybersecurity committee and referred to Senate Ways and Means) has not been enacted. If enacted, it would do the following. SECTION 1 would add G.L. c. 29, section 2GGGGGG, creating the Massachusetts Artificial Intelligence Innovation Trust Fund. The secretary of economic development would be trustee. The secretary would use the fund, in consultation with the Massachusetts Technology Park Corporation, for grants to AI companies in key sectors, AI entrepreneurship programs, and AI research grants. Money in the fund could be spent without further appropriation and would not revert to the General Fund. SECTION 2 would add a new G.L. chapter 93M, the "Transparency in Frontier Artificial Intelligence Act." It would cover "frontier developers" who train foundation models with more than 10^26 operations of computing power. "Large frontier developers" are those with more than $500 million in annual revenue. Large developers would have to write, publish, follow, and at least annually review a frontier AI framework for managing catastrophic risk. Catastrophic risk means a risk of more than 50 deaths or serious injuries, or more than $1 billion in property damage, from a single incident. All frontier developers would have to publish transparency reports when they deploy a new or substantially modified model. Large developers would also have to include summaries of their catastrophic-risk assessments in those reports. Large developers would have to send the attorney general summaries of internal-use risk assessments. Developers could not make materially false or misleading statements about catastrophic risk, with a good-faith exception. Redactions would be allowed to protect trade secrets, security, and similar interests. Developers would have to report critical safety incidents to the attorney general within 15 days, or within 24 hours to appropriate authorities if there is an imminent risk of death or serious injury. The attorney general would set up reporting mechanisms and issue annual anonymized reports starting Jan. 1, 2027. The attorney general could designate federal standards that developers may comply with instead. These incident reports would be exempt from the public records law (chapter 66). The attorney general would also assess and recommend updates to the definitions each year. Large developers that fail to comply could face civil penalties of up to $1,000,000 per violation, enforced only by the attorney general (Section 4). Section 7 would protect "covered employees" who report catastrophic-risk dangers or violations from retaliation. It would require notice of their rights, an anonymous internal disclosure process at large developers, attorney's fees, a burden shift to the developer, and injunctive relief. Section 6 would create a 14-member consortium within the Executive Office of Technology Services and Security. It would develop a framework for "MassCompute," a public cloud computing cluster, subject to appropriation. Section 9 would let the attorney general issue regulations, in consultation with MassCompute.

← Back to the tracker