(BOSTON 10/21/2025) — Last week, the Senate Committee on Advanced Information Technology, the Internet and Cybersecurity voted to advanced five pieces of legislation aimed at placing guardrails around rapidly-advancing artificial intelligence technology. The bills include policies that create broad operational guidelines and responsibilities for those developing and using AI, as well as more targeted actions that will limit the use of AI for the creation of child sexual abuse material or election misinformation, for example. Members of the Committee voted unanimously on a bipartisan basis to report the bills with a favorable recommendation.
“As things stand, there are very few legal limitations for what individuals, corporations, and developers can do with artificial intelligence technologies like large language models or generative AI. There are even fewer legal protections for those whose data is used to train these models. This regulatory void leads to a lot of uncertainty for everyone, whether you use AI or not,” said Chair of the Senate Committee on Advanced Information Technology, the Internet and Cybersecurity Senator Michael Moore (D-Millbury). “The five bills we are advancing out of Committee this week represent a step toward reasonable, common-sense regulations that will protect Bay Staters from the negative effects of AI while maintaining Massachusetts’ status as one of the most innovative economies in the world. Striking this balance is critically important, and thanks to the work of my colleagues on the Committee, I believe we are doing just that.”
The bills include:
An Act fostering artificial intelligence responsibility – S.35
This bill defines the term ‘electronic monitoring tool’ in the workplace as a device or system that collects data related to worker activity or communication, and creates narrow guidelines for how these tools can be used. Specifically, the legislation allows the use of electronic monitoring tools to ensure the quality of goods or services, conduct assessments of worker performance, ensure compliance with labor laws, protect the health and safety of workers, and administer wages; however the language also specifies that employers may only use these tools when employees are on-duty and employees must give their written consent before they may be electronically monitored.
S.35 specifies that while electronic monitoring tools may be used in the workplace, they may not rely primarily on data collected through these systems when making hiring, promotion, and disciplinary decisions. If an employer wishes to use data collected through these systems when making these types of decisions, they must disclose its use to the affected employee no less than 30 days prior to the decision going into effect.
In an effort to protect employees, the bill also limits the sale or transfer of employee data as well as requires that employee data is maintained in compliance with Massachusetts’ cybersecurity and privacy laws. Further, this legislation requires employers conduct impact assessments and evaluations for discriminatory outcomes related to electronic monitoring tools.
An Act promoting economic development with emerging artificial intelligence models and safety – S.2630
S.37’s primary objective is the creation of a permanent board called MassCompute which will be tasked with the development and deployment of artificial intelligence that is safe, ethical, equitable, and sustainable. MassCompute will be made up of experts and stakeholders in both private industry and public service, and in partnership with the Massachusetts Attorney General, will create and implement regulations related to AI and its usage.
To promote AI development in Massachusetts, the legislation creates an Artificial Intelligence Innovation Trust Fund to support the work of MassTech’s AI Hub, as well as directs the Fund to establish and promote AI entrepreneurship programs.
This legislation is modeled after California legislation which was recently signed into law by Governor Newsom.
An Act to protect against election misinformation – S.2631
This bill prohibits a person or political committee from maliciously distributing deceptive election-related information, AI generated or not, with the intent to mislead voters within 90 days of an election as to:
The date, time, or place of an election
The requirements, methods, and deadlines for registering to vote
The requirements, methods, and deadlines for voting
Any certification related to an election
The express endorsement of a candidate or ballot initiative by a political party, elected official nonprofit organization, or other person
S.44 maintains exceptions to this law for news media so long that the broadcast clearly acknowledges that the media is manipulated or there are questions about its authenticity. It also maintains exceptions for 1st Amendment-protected satire or parody.
An Act relative to the use of artificial intelligence and other software tools in healthcare decision-making – S.2632
Modeled after a California law, this legislation ensures that decisions about medical treatments are made by health care providers, and not solely determined by artificial intelligence algorithms used by health insurers. While the bill does not prohibit the use of AI in assisting with decision making, it includes protections against bias and errors in the decision making process.
Further, the bill limits the use of AI in therapy by ensuring therapy or psychotherapy services may only be conducted by a licensed professional. AI as “supplementary support” may be used so long as the patient has provided written consent. These limits on the use of AI in therapy are modeled after an Illinois law.
An Act relative to AI-generated child sexual abuse material – S.2633
S.48 expands existing protections against the creation and distribution of child sexual abuse material (CSAM) to apply to CSAM created in whole or in part through the use of generative AI. It also defines “child sexual abuse visual materials” as images, including computer generated images and including those made through digitization, that appear to a reasonable person as an authentic representation of a minor.
Having been advanced out of the Senate Committee on Advanced Information Technology, the Internet and Cybersecurity with favorable recommendations, the bills now advance for further consideration in the Senate.
###