Judiciary Committee approves substitute for HB60 to regulate high‑risk predictive AI; original bill not advanced

Chair Chandler presented House Bill 60 and told the committee the measure would create a state framework for so‑called high‑risk predictive AI systems used to make consequential decisions about employment, housing, education and lending. “This bill…provides a framework for companies that create, develop AI that is predictive,” Chair Chandler said, adding the draft places duties both on developers and on organizations that deploy software in decisions that materially affect people.

The sponsor and experts described key elements of the committee substitute: a definition of “consequential decision” limited to actions that have a material, legal or similar significant effect on consumers (including employment, housing, education and lending); a definition of “high‑risk artificial intelligence system” tied to systems that substantially factor into consequential decisions; duties of reasonable care for developers and deployers; annual or event‑driven risk and impact assessments from deployers; notice and appeal rights for people denied opportunities; and reporting or notification obligations to the attorney general on identified risk incidents.

Experts who joined the sponsor to explain technical details included Steve Wimmer of the Transparency Coalition and Chris Moore of the Santa Fe Institute. Moore emphasized the bill’s scope: it excludes generative, conversational systems and focuses on predictive systems that materially affect people.

The committee heard more than an hour of public testimony in both opposition and support. Insurance industry trade associations—represented by Brent Moore (American Property Casualty Insurance Association and others) and Ward Tisdale (National Association of Mutual Insurance Companies)—urged carve‑outs for insurers, saying existing insurance regulation and supervisory regimes create duplicative and costly oversight. PNM and several industry groups asked for further study or a narrower approach. Opponents warned that extensive disclosure requirements could expose trade secrets and create practicability concerns.

Supporters included Lena Weber (ACLU of New Mexico), Kiara (Equality New Mexico), Melanie Moses (University of New Mexico; New Mexico AI Consortium), Grace Giddey (Consumer Reports) and representatives of worker and consumer groups. Supporters emphasized the bill’s transparency provisions and its protections against algorithmic discrimination in core domains.

Committee members asked technical questions about the bill’s definitions (developer, deployer, substantial factor, consequential decision), the scope of the notice and appeal provisions, timing for a required cure by a deployer, and whether the statute would create a private right of action. The sponsor and experts said the substitute narrows and clarifies several definitions from the first draft, that deployers must file impact assessments annually and on significant change, and that a limited private right of action in the substitute is narrowly framed (injunctive relief and attorney’s fees but not compensatory damages). The attorney general’s office was described in testimony as an enforcement backstop.

After discussion the committee voted on the motion recorded at today’s session: the committee did not recommend passage of the original bill as drafted, but it recommended passage of the committee substitute. The chair announced the committee outcome: "do not pass on House Bill 60, but do pass on committee substitute 60." The substitute will move forward according to legislative procedures.

Why it matters: HB60 addresses an expanding set of commercial and institutional uses of automated decision systems. The substitute attempts to balance consumer and civil‑rights protections (transparency, contestability, assessment and notice) with carve‑outs and clarifications sought by industry and academic stakeholders. Because the substitute creates reporting obligations and a role for the attorney general, passage as eventually enrolled would affect how private companies document and deploy predictive models that substantially influence housing, employment, credit and other essential services.

What’s next: The committee’s do‑pass recommendation for the committee substitute advances that text to the next step in the House process. Further amendments are possible in subsequent committee or floor action; several industry witnesses said they will continue negotiations with the sponsor.