Operators must submit an annual report to the Attorney General containing: (1) the total number of times the crisis intervention protocol was triggered during the preceding calendar year, and (2) a summary of the results of the most recent biennial compliance audit.

The commission must submit an annual public report by December 31 to the Governor, the legislative clerks, and the joint committee on state administration and regulatory oversight, detailing the extent of government automated decision-making, findings, recommendations for regulatory or legislative action (including areas where state agencies should not use automated decision systems), implementation timelines, cost estimates, and progress on prior recommendations.

The Department of State Police must document each facial recognition search request and search performed as a public record and report this information quarterly to the Executive Office of Public Safety and Security, including date, time, system used, offenses investigated, match counts, requestor identity and agency, warrant or emergency documentation, and presumed race and gender of probe image subjects.

Non-law enforcement public agencies must document each facial recognition search requested or performed by their officials as a public record and report this information quarterly to the Executive Office of Public Safety and Security, including date, time, requestor identity, reason for search, search operator identity, system used, match counts, and presumed race and gender of probe image subjects.

The department of state police must document every facial recognition search request and search performed as a public record, and must report this information quarterly to the executive office of public safety and security, including date/time, system used, offenses under investigation, match counts, requester identity, warrant or emergency documentation, and the presumed race and gender of probe image subjects.

Each non-law-enforcement public agency must document every facial recognition search requested and performed by its public officials as a public record, and must report this information quarterly to the executive office of public safety and security, including date/time, requester identity, reason, system used, match counts, and the presumed race and gender of probe image subjects.

The Department of State Police must document each facial recognition search request and search performed as a public record and report quarterly to the Executive Office of Public Safety and Security, including date, time, system used, offenses investigated, matches returned, requesting officer identity and agency, warrant or emergency documentation, and the presumed race and gender of the probe image subject.

Each non-law-enforcement public agency must document as a public record every facial recognition search requested or performed by its officials and report quarterly to the Executive Office of Public Safety and Security, including date, time, requesting individual, reason, conducting individual, system used, matches returned, and the presumed race and gender of the probe image subject.

Health insurance carriers must submit quarterly reports to the Commissioner on adverse decisions and grievances, now disaggregated by zip code, race, ethnicity, gender, and age, and including: the number and outcomes of grievances, grievance monetary values, average hold and call times for grievance and appeal call centers, and the monetary value of adverse decision cases outside the subtitle's scope.

Law enforcement agencies must prepare and publish by February 1 each year an annual report covering the prior calendar year, disclosing (1) each facial recognition system name and databases searched, (2) total searches per system with associated crime types, (3) total matches leading to further investigative action per system and database, and (4) any data breaches or unauthorized uses.

Law enforcement agencies must prepare and publish by February 1 each year an annual report covering the prior calendar year's facial recognition activity, including total searches performed and associated crime types, matches returned, data breaches or unauthorized uses, extent of use, purpose descriptions including sole-basis versus supporting use, intended benefits with supporting data, technology capabilities and limitations, data inputs, result types, and any known or suspected policy violations.

The Governor's Office of Crime Prevention, Youth, and Victim Services must submit to the Governor and the General Assembly by October 1 each year a consolidated report of all law enforcement agency facial recognition reports, disaggregated by agency.

Carriers must report quarterly to the Bureau of Insurance, beginning April 15, 2026, on the number of AI-related coverage denials, the outcome of physician reviews, and the number and outcome of appeals of AI-related determinations.

The Minnesota Supreme Court must annually submit to the legislature and make publicly available a comprehensive report summarizing all facial recognition surveillance activity, including the number of applications and orders, details of each order, frequency and scope of use, resources expended, and the number of misidentifications including arrests not resulting in charges.

The Minnesota Supreme Court must annually compile and submit to the legislature — and make publicly available — a comprehensive report on government use of facial recognition technology, including the number of applications, grants, denials, duration of orders, applying agencies, facility types, frequency of use, number of persons analyzed, resources expended, and misidentification counts.

Local law enforcement agencies that used facial recognition technology must submit an annual report by June 30 to the criminal intelligence information section covering the number of searches run, offenses investigated, and arrests and convictions resulting from the searches for the prior calendar year.

The Department of Justice must submit an annual report by September 1 to the economic affairs and law and justice interim committees, consolidating local law enforcement facial recognition data and its own usage data including requesting agencies, search counts, offenses investigated, and resulting arrests and convictions.

Third-party vendors providing facial recognition services to state agencies must submit annual reports by June 30 to thecontracting agency covering warrants/subpoenas/court orders received and a vendor audit summary. The state agency must forward these reports to the legislative interim committees and the information technology board by September 1.

The Department must submit an annual report by January 15 to the Environmental Review Commission, the Joint Legislative Oversight Committee on Agriculture and Natural and Economic Resources, and the Fiscal Research Division, covering: (1) number and type of permits where AI was used, (2) AI system performance assessments, (3) processing-time comparisons, (4) AI error incidents and corrective actions, and (5) recommendations on continuation, expansion, modification, or discontinuation of AI use.

Licensees must implement continuous monitoring systems for safety and risk indicators and submit quarterly performance reports to the Department, including incident reports.

Licensees must implement continuous monitoring systems for safety and risk indicators and submit quarterly performance reports, including incident reports, to the Department.

Employers who receive the Department of Employment, Training and Rehabilitation's AI job displacement survey must complete and return it. The Department must compile results into an anonymized annual report and submit it to the Governor and Legislature by December 1 each year.

Deployers must compile aggregated feedback data into an annual performance report and submit it to the AI Council by January 1 of each year.

The Department of State must submit an annual report to the Commerce Committees of both chambers by January 1 of each year, including the total number of registered AI businesses, the number of registrants received by each state agency, and expenses incurred by the Department.

The Department of State must submit annual reports to the Senate and House Commerce Committee leadership by January 1 of each year, disclosing the total number of registered AI businesses, the number of registrants received by each state agency, and expenses incurred by the Department.

Non-small-business online platforms that engage in content moderation must publish at least annually a publicly accessible, machine-readable transparency report disclosing total content moderation decisions, breakdowns by policy type, coordinated-behavior responses, demographic data, human vs. automated moderation, appeal rates and reversal rates, government demands, and other FTC-specified information. The report must be available without requiring account creation and must allow free copying and reuse of the data.

Non-small-business online platforms that engage in content moderation must publish at least annually a publicly available, machine-readable transparency report detailing total and disaggregated content moderation decisions broken down by policy category, coordinated-behavior responses, user demographics, moderation method (automated, employee, external), appeal and reversal rates, government-initiated moderation requests (by jurisdiction), and decision types.

State judges and prosecutors must annually report to a designated state agency detailed information about facial recognition orders issued, searches run, offenses investigated, arrests and convictions resulting from searches (disaggregated by race, ethnicity, gender, and age), motions to suppress, and databases used.

Federal judges and prosecutors must annually submit to the Director of the Administrative Office of the United States Courts the same categories of facial recognition use data required of state counterparts.

Operators must publish a quarterly transparency report, in a machine-readable open format with an open license and in an easily accessible location, disclosing total and algorithm-driven political-email spam flagging counts and percentages broken down by Republican and Democratic party affiliation, plus a summary of the tools and techniques used to determine which political campaign emails to flag.

Non-small-business online platforms that engage in content moderation must publish at least annually a public transparency report including total content moderation decisions, breakdowns by policy category, coordinated-behavior response, aggregate demographics, moderation method, appeal and reversal rates, government demands, and decision types. Reports must be publicly accessible without account creation, in machine-readable format, freely copyable, and available in all service languages.

Non-small-business online platforms that engage in content moderation must publish at least annually a publicly accessible transparency report containing total content moderation decisions, breakdowns by policy category, coordinated-behavior responses, aggregate demographics, moderation method, appeal and reversal rates, government requests, and decision types. Reports must be available without account creation, in all supported languages, and in machine-readable format with freely copyable data.

Suppliers must generate aggregated user engagement estimates — calculated separately for minors and adults using a statistically valid methodology — identifying median and 90th-percentile hours of use and number of chatbot interactions for each group.

Deployers of high-risk AI systems must submit one-month, six-month, and twelve-month post-deployment testing results to the Division of Artificial Intelligence in the first year of deployment, showing system reliability, variance over the testing periods, and mitigation strategies for any variances.

Beginning January 1, 2023, each agency using an automated decision system must publish on its website annual metrics on the number of requests for human review of decisions rendered by the system and the outcomes of those reviews.

The Algorithmic Accountability Review Board must conduct an annual review of agency audits beginning January 1, 2026, and compile a public report covering agency compliance, known violations, systematic bias issues, and recommendations, published on the Office's website by March 1 annually (first report by March 1, 2025).

Beginning January 1, 2025, each agency using an automated decision system must publish on its website annual metrics on the number of requests for human review of automated decisions and the outcomes of those reviews.

The Artificial Intelligence Advisory Council must submit quarterly reports to the Governor, the President of the Senate, and the Speaker of the House covering findings from agency AI inventory reports, national AI policy recommendations, an assessment of AI's impact on residents' rights and interests, and policy recommendations for privacy protection, anti-discrimination, and AI development promotion.

The commission must submit an annual public report by December 31 to the Governor, the legislative clerks, and the joint committee on state administration and regulatory oversight, detailing the extent of government automated decision-making, findings, recommendations for regulatory or legislative action (including areas where state agencies should not use automated decision systems), implementation timelines, cost estimates, and progress on prior recommendations.

Law enforcement agencies must prepare and publish by February 1 each year an annual report covering the prior calendar year's facial recognition activity, including total searches performed and associated crime types, matches returned, data breaches or unauthorized uses, extent of use, purpose descriptions including sole-basis versus supporting use, intended benefits with supporting data, technology capabilities and limitations, data inputs, result types, and any known or suspected policy violations.

Non-small-business online platforms that engage in content moderation must publish at least annually a publicly accessible, machine-readable transparency report disclosing total content moderation decisions, breakdowns by policy type, coordinated-behavior responses, demographic data, human vs. automated moderation, appeal rates and reversal rates, government demands, and other FTC-specified information. The report must be available without requiring account creation and must allow free copying and reuse of the data.

Non-small-business online platforms that engage in content moderation must publish at least annually a publicly available, machine-readable transparency report detailing total and disaggregated content moderation decisions broken down by policy category, coordinated-behavior responses, user demographics, moderation method (automated, employee, external), appeal and reversal rates, government-initiated moderation requests (by jurisdiction), and decision types.

The Director of OMB must submit to Congress a report on agency implementation of and conformity to the NIST AI Risk Management Framework within one year of enactment and at least every three years thereafter.

Operators must publish a quarterly transparency report, in a machine-readable open format with an open license and in an easily accessible location, disclosing total and algorithm-driven political-email spam flagging counts and percentages broken down by Republican and Democratic party affiliation, plus a summary of the tools and techniques used to determine which political campaign emails to flag.

Non-small-business online platforms that engage in content moderation must publish at least annually a public transparency report including total content moderation decisions, breakdowns by policy category, coordinated-behavior response, aggregate demographics, moderation method, appeal and reversal rates, government demands, and decision types. Reports must be publicly accessible without account creation, in machine-readable format, freely copyable, and available in all service languages.

Non-small-business online platforms that engage in content moderation must publish at least annually a publicly accessible transparency report containing total content moderation decisions, breakdowns by policy category, coordinated-behavior responses, aggregate demographics, moderation method, appeal and reversal rates, government requests, and decision types. Reports must be available without account creation, in all supported languages, and in machine-readable format with freely copyable data.