AI Frontier Institute

Online platforms required to provide information pertaining to algorithm use, design transparency and user choice required, civil penalties provided, and rulemaking authorized.

MNHF3980In committee
Published automatically

View primary source →

H.F. No. 3980 has not been enacted. It was introduced 03/05/2026 and referred to the House Committee on Commerce Finance and Policy. If enacted, it would add Minnesota Statutes section 325M.35, "Online Platform Metrics." It would apply to "covered online platforms" that do business in Minnesota and use algorithmic recommender systems that rely at least partly on users' personal data. Narrow exceptions cover user-selected settings, device information, and certain search queries. Under Subd. 3, these platforms would have to prominently disclose each recommender system, its inputs and data sources, and its weights grouped into four quartiles by importance. They would also have to disclose annually the objectives, key results, and metrics used to evaluate the product teams that design these systems. The commissioner of commerce would adopt rules clarifying the disclosures. Under Subd. 4, recommender systems that use personal data must by default maximize "long-term user value" metrics. Users would get an accessible interface to state which item types they want recommended or blocked, and the platform must take reasonable steps to honor those choices. The platform could not withhold, degrade, or raise the price of a service because a user exercised these rights. Under Subd. 5, systems serving covered minors must default to maximizing long-term user value metrics applicable to minors. Under Subd. 6, platforms must keep at least one holdout group and subject every design change to a long-term (at least 12 months) holdout assessment. They must publish annual disclosures of the metrics and anonymized aggregate results for the holdout group and for other users. They must also obtain an annual independent audit at their own expense. The commissioner would write rules for these assessments and could exempt design changes that reduce or prevent direct and immediate harm without increasing engagement or revenue. Under Subd. 7, a violation would be an unfair and deceptive act under section 325D.44, and the attorney general could enforce it under section 8.31. Users could recover remedies under section 325D.45. They could also recover $5,000 per user per violation (CPI-adjusted) or actual damages, whichever is greater. For reckless or knowing violations they could recover punitive damages of $7,000 per violation (CPI-adjusted) or actual damages, whichever is greater. They could also recover attorney fees and injunctive or declaratory relief. Subd. 2 states that the section adds to other Minnesota law, that the law giving the greatest consumer protection controls in a conflict, and that it must be read consistently with the First Amendment and 47 U.S.C. section 230.

Status history

Current status as of 2026-03-05

  1. In committee

    2026-03-05

    observed 2026-09-02

Impact areas

← Back to the tracker