🗺️ AI Laws USA

8 AI policy moves you should know about — week of July 7, 2026

By AI Laws USA ·

Illinois signed the nation's first frontier-AI audit law; Washington's employer-monitoring prohibition went live; and Nashville packed a council chamber with 200 speakers — all in one week.

AI Laws USA tracks every U.S. federal, state, county, and city AI law on a single live map at ailawsusa.com. The week of July 7, 2026 was unusually dense: a governor signed the first U.S. law requiring independent audits of the largest AI models; another state's landmark employer-monitoring law went into effect; and the Nashville data-center battle graduated from first reading to a record-setting public hearing. Here are the eight moves that mattered most.

  1. NEW LAW

    1. Illinois becomes the first state to require independent third-party audits of frontier AI models

    Illinois (state — SB 315) · Signed July 6, 2026 by Governor Pritzker — effective January 1, 2027 (audit requirements operative January 1, 2028)

    Governor Pritzker signed SB 315, the Artificial Intelligence Safety Measures Act, on July 6, 2026 — the first U.S. law to mandate independent third-party safety audits of the largest AI developers (those with >$500M revenue training above defined compute thresholds). Covered developers must publish a frontier AI safety framework addressing catastrophic risks (50+ deaths or $1B+ in damages), file pre-deployment transparency reports, report critical safety incidents within 72 hours, and protect whistleblowers. The House passed it 110-0; the Senate 52-5. The only U.S. frontier-AI law with a mandatory audit requirement.

    Cite this source ↗ · View live entry on the map →

  2. NEW LAW

    2. Washington's employer-monitoring prohibition — the strictest in the country — took effect July 1

    Washington (state — SHB 1672) · Effective July 1, 2026

    Washington's Substitute House Bill 1672 went live July 1, 2026, giving Washington employees the strongest workplace-monitoring rights of any U.S. state. Employers must give 15 days' written notice before monitoring starts or changes; off-duty and private-space monitoring is banned outright; AI-based emotion recognition, gait recognition, and facial recognition are prohibited in employment decisions; impact assessments are required before deploying automated decision systems; and human oversight is required for ADS-driven performance evaluations. Enforcement runs through both a private right of action ($500+ per violation plus attorney fees) and the Department of Labor & Industries (up to $10,000 per violation).

    Cite this source ↗ · View live entry on the map →

  3. NEW LAW

    3. Maryland enacted an employer ADS disclosure and impact-assessment law, effective October 1

    Maryland (state — HB 1339) · Signed April 28, 2026 — effective October 1, 2026

    Maryland HB 1339, signed by Governor Moore on April 28, 2026, requires employers to give applicants and employees written notice before using any automated decision system in hiring, promotion, or termination; to conduct annual disparate-impact assessments for each ADS; and to retain records for three years. Civil penalties reach $10,000 per violation per day. It takes effect October 1, 2026, giving Maryland the most detailed ADS employer-disclosure regime after Illinois HB 3773 and New York City's Local Law 144.

    Cite this source ↗ · View live entry on the map →

  4. NEW LAW

    4. Maine's employer surveillance disclosure law takes effect this week

    Maine (state — LD 61 / P.L. 2025, Ch. 524) · In effect ~July 14, 2026 (became law January 11, 2026)

    Maine LD 61 (P.L. 2025, Ch. 524) — which became law without Governor Mills's signature on January 11, 2026 — takes effect approximately July 14, 2026, about 90 days after the legislature adjourned. Employers must give advance written notice before any monitoring begins, disclose surveillance practices to job applicants, provide annual notices to current employees, and allow workers to refuse monitoring software on their personal devices. Audiovisual monitoring of employees' homes, personal vehicles, or personal property is prohibited. Civil fines of $100–$500 per violation. Maine joins Connecticut, New York, and Delaware in the Northeast's cluster of workplace-monitoring disclosure laws.

    Cite this source ↗ · View live entry on the map →

  5. NEW LAW

    5. The Pacific Northwest now has two companion chatbot safety laws — and Oregon's bites hardest

    Oregon (state — SB 1546); Washington (state — HB 2225) · Oregon: signed March 31, 2026; Washington: signed March 24, 2026 — both effective January 1, 2027

    Oregon and Washington each enacted AI companion chatbot safety laws in March 2026, creating a coordinated regional framework. Both require recurring disclosures (every 3 hours for adults, every hour for minors), prohibit manipulative mechanics targeting minors, and mandate crisis protocols for suicidal ideation. The key difference: Oregon SB 1546 includes a direct private right of action ($1,000 per violation plus actual damages and attorney fees) — making it the first U.S. chatbot law where any harmed person can sue directly. Washington HB 2225 enforces through the AG under the Consumer Protection Act. Both take effect January 1, 2027.

    Cite this source ↗ · View live entry on the map →

  6. PROPOSED

    6. Nashville's data-center moratorium survived the biggest public hearing in Metro Council history

    Nashville, TN (city — Metro Council) · Passed second of three required Council readings — third reading July 21, 2026

    On July 7–8, 2026, the Nashville Metro Council held a marathon overnight public hearing in which more than 200 residents spoke in support of the proposed data-center moratorium — the largest public-comment turnout in Metro Council history. Both the temporary moratorium (pausing all approvals until November) and a companion permanent-restrictions bill (banning data centers over 500,000 square feet, limiting smaller facilities to heavy-industrial zoning with special approval) passed their second readings unanimously. The third and final vote is scheduled for July 21. The bills would transform Nashville/Davidson County from a permissive data-center market into one of the more restrictive major-city jurisdictions in the South.

    Cite this source ↗ · View live entry on the map →

  7. NEW LAW

    7. Two more cities banned algorithmic rent-setting software: Hoboken, NJ and Santa Monica, CA

    Hoboken, NJ (city); Santa Monica, CA (city) · Hoboken: effective July 9, 2025 (tracked this week); Santa Monica: effective ~June 24, 2025

    Hoboken's City Council voted 8-0 on July 9, 2025 to ban algorithmic rent-fixing software — software that pools nonpublic competitor pricing data to coordinate rents across multiple properties (targeting RealPage YieldStar-type systems). Penalties reach $20,000 per violation. Santa Monica passed a similar ban in late June 2025, with an added tenant affirmative defense in eviction proceedings — motivated in part by the January 2025 LA wildfire housing crisis. Both cities join a growing list that includes Seattle, San Diego, Minneapolis, Berkeley, Providence, and Jersey City. New Jersey's statewide FAIR Act is still awaiting Governor Sherrill's signature.

    Cite this source ↗ · View live entry on the map →

  8. NEW LAW

    8. New York's insurance AI rule is the strongest in the country — and it predates this week

    New York (state — NYDFS Insurance CL No. 7) · In effect since July 11, 2024 — newly indexed this week

    AI Laws USA newly indexed New York's NYDFS Insurance Circular Letter No. 7 (2024) — issued July 11, 2024 — which goes further than the NAIC Model Bulletin adopted by 24+ other states. Where the NAIC bulletin sets governance expectations, NYDFS CL No. 7 prohibits any insurer from using an AI system or external data source in underwriting or pricing unless it first completes a 'proxy assessment' demonstrating the system does not produce unfair or unlawful discrimination against protected classes. When an AI-influenced adverse decision is made, the insurer must notify the applicant in writing within 15 days. Enforceable through DFS market-conduct examinations across all lines of insurance.

    Cite this source ↗ · View live entry on the map →

The thread connecting this week's moves: policymakers at every level are trying to hold specific harms accountable before they scale. Illinois's audit requirement targets catastrophic risk in frontier models before deployment. Washington's monitoring law targets emotion AI and off-duty surveillance before it becomes normalized in workplaces. Nashville's hearing packed a council chamber because residents saw data centers as a fait accompli unless they showed up in force — and they did. Track every one of these as they move, plus every law that's already in effect for any U.S. address, at ailawsusa.com. New posts every Monday.

All posts