AI Frontier Institute

To ensure accountability and transparency in artificial intelligence systems

View primary source →

House Bill No. 94 (Massachusetts, 194th General Court) has not been enacted. It was referred to the Joint Committee on Advanced Information Technology, the Internet and Cybersecurity, and the record shows it is still at the committee stage. If enacted, it would create a new Chapter 93M of the General Laws, "Artificial Intelligence Accountability and Consumer Protection." The chapter would cover algorithmic discrimination by AI systems, with extra duties for "high-risk" systems that materially influence consequential decisions in education, employment, finance or lending, housing, healthcare, insurance, and legal or government services. Developers (Chapter 93M, Section 2) would have to: - use reasonable care to identify, mitigate, and disclose risks of algorithmic discrimination; - give deployers documentation on intended uses, known limitations and risks, and training data; - notify the Attorney General and deployers of known or foreseeable discrimination risks within 90 days of discovery; - publish a plain-language public statement on their website. Deployers of high-risk systems (Section 3) would have to: - maintain a risk management program aligned with standards such as the NIST AI Risk Management Framework; - complete annual impact assessments, updated after any substantial modification, with state-provided templates; - notify consumers when AI materially influences a consequential decision, explain how it did so, and offer a process to appeal or correct adverse decisions; - publicly disclose the types of high-risk systems they use and their risk mitigation strategies. Corporate disclosure (Section 4): corporations that use AI to target consumer groups or influence behavior would have to disclose their purposes, methods, behavioral influence techniques, and third-party partners. They would post this on their websites and include it in terms and conditions. Proprietary information would be protected from public disclosure. Consumers would be notified when AI targets or influences them in a way that materially affects their decisions, or when algorithms set pricing, eligibility, or access to services. Exemptions (Section 5): - businesses with fewer than 50 employees that do not use proprietary data to train AI; - low-risk procedural systems; - entities subject to equivalent or stricter federal AI regulation. Enforcement (Section 6): - The Attorney General would have exclusive enforcement authority, and violations would be unfair or deceptive practices under Chapter 93A. - Affirmative defenses would be available for violations that are remedied or for compliance with recognized standards. - There would be no private right of action. Other provisions: - Section 7 would let the Attorney General issue rules on documentation, impact assessments, risk management standards, consumer notifications, and recognized frameworks. - Section 8 would require a public education campaign. - Under Sections 9–11, Sections 1, 4, 5, and 8 and the Chapter 93A amendment would take effect 180 days after passage. Sections 2, 3, 6, and 7 would take effect one year after passage.

← Back to the tracker