Massachusetts is poised to become the fourth state to regulate the most powerful artificial intelligence models, requiring companies to disclose risks of catastrophic harm. State Senator Barry Finegold, a Democrat who chairs the Committee on Economic Development and Emerging Technologies, sponsored the Transparency in Frontier Artificial Intelligence Act. The bill is wrapped into a broader economic development package that the state Senate passed last week and is a priority for Democratic Governor Maura Healey.
The legislation targets the largest, most capable AI systems - often called frontier models - and forces their developers to publicly report safety tests, potential misuse scenarios, and mitigation plans. The goal is to guard against harms such as autonomous cyberattacks, bioweapon design, or large-scale societal disruption before those models are widely deployed.
What the bill requires
Under the proposed act, companies training models above a defined computing-power threshold must submit transparency reports to state regulators. Those reports would detail the model's intended use, known limitations, and results of adversarial testing. Companies that fail to comply could face penalties.
Finegold's bill mirrors emerging laws in California, Connecticut, and Washington State, which have each passed varying forms of frontier AI transparency requirements in the past two years. The Massachusetts version emphasizes public disclosure rather than government licensing, a distinction that may shape future federal debates.
Why regulators are acting now
Lawmakers cite the rapid pace of AI capability growth and the limited external visibility into how leading labs test their most advanced models. The bill does not ban any technology but forces companies to share what they know - and what they don't know - about catastrophic failure modes.
For government employees who work with or oversee AI procurement, this law would create enforceable standards for vendor accountability. Training and policy resources, such as the AI for Government tag page, and the AI Learning Path for Policy Makers, can help officials understand these requirements and apply them in their own agencies.
Why this matters for government professionals
If the bill becomes law, state and local government employees in Massachusetts - and eventually elsewhere - will need to evaluate AI vendors against mandated transparency reports. That shifts due diligence from trusting a company's public statements to verifying documented risk disclosures. For federal workers, the Massachusetts approach is a signal: transparency mandates are gaining momentum, and similar requirements could appear in national procurement rules. Government professionals who understand frontier AI risk categories and reporting standards will be better positioned to write contracts, audit compliance, and protect public safety.
Your membership also unlocks:
