Trackers · Living reference

AI Regulation Tracker

Every AI law and regulation that governs how professionals can deploy AI — EU AI Act, US state hiring laws, FDA and SEC actions — with effective dates and primary sources. The running reference for compliance, legal, HR, and finance teams.

Fast-moving items are provisional. The EU Digital Omnibus (high-risk deadline deferrals) and the federal executive order on state-law preemption are agreed in principle but not yet settled law — the status column flags what binds today versus what is contested.

Last updated 2026-07-16 · 16 measures

JurisdictionLawDomainStatusEffectiveSource
EUAI Act — Article 50 transparency dutiesNOT affected by the high-risk deferral: providers and deployers must disclose when a user is interacting with an AI system and label AI-generated or -manipulated content (deepfakes, synthetic audio/text). Still takes effect Aug 2, 2026.GeneralIn forceAug 2, 2026Covington (Inside Privacy)insideprivacy.com
CaliforniaAI Transparency Act (SB 942, amended by AB 853)Large generative-AI providers (>1M monthly users) must offer a free AI-detection tool and attach visible ("manifest") plus embedded ("latent") disclosures to AI-generated content. AB 853 (signed Oct 13, 2025) moved the operative date from Jan 1, 2026 to Aug 2, 2026 to align with the EU AI Act; large-platform and GenAI-hosting duties phase in Jan 1, 2027.GeneralIn forceAug 2, 2026Mayer Brownmayerbrown.com
US FederalBIS Final Rule — Enhanced Favorable Treatment for the UAE (RIN 0694-AK54)Reclassifies the UAE to Country Group A:5 and creates Supplement No. 8 to Part 740 listing approved UAE government agencies and commercial entities (including G42, Core42, and UAE subsidiaries of Amazon, Apple, Google, Meta, Microsoft, OpenAI, Oracle and X.AI) that may receive advanced AI computing items and use License Exception STA license-free; other UAE recipients still require licenses and must request a BIS advisory opinion. G42 and Core42's eligibility expires 270 days from July 10, 2026 unless they become US companies.GeneralIn forceJul 10, 2026Federal Register / BIS Final Rule (Justia mirror)regulations.justia.com
IllinoisHB 3773 (amends Human Rights Act)Makes discriminatory AI in hiring a civil-rights violation — reaching unintentional disparate impact — bars zip-code proxies, and requires applicant notice when AI is used in covered employment decisions.EmploymentIn forceJan 1, 2026Hinshaw & Culbertsonhinshawlaw.com
TexasResponsible AI Governance Act (HB 149 / TRAIGA)Bars developing or deploying AI with intent to unlawfully discriminate (intent-based — disparate impact alone is not a violation). AG-only enforcement, 60-day cure period, no private right of action.GeneralIn forceJan 1, 2026Norton Rose Fulbrightnortonrosefulbright.com
New York CityNYC Bar Formal Opinion 2025-6 — AI Recording, Transcription and Summarization of Client ConversationsLawyers using AI tools to record, transcribe or summarize client conversations must obtain informed client consent; undisclosed AI recording is inconsistent with candor duties and Rule 8.4, and lawyers must assess vendor confidentiality, biometric-data laws, all-party consent statutes, and tactical implications before enabling AI notetakers. Duty of competence (Rule 1.1) requires understanding how the tool transmits, retains, and reuses data.GeneralIn forceDec 2025New York City Bar Associationnycbar.org
CaliforniaADMT employment-discrimination regulationsCalifornia Civil Rights Council regulations on automated-decision-making technology (ADMT) in employment decisions take effect Oct 1, 2025: ADMT that causes disparate-impact discrimination can violate state law, four-year record retention applies, and liability extends to an employer’s agents.EmploymentIn forceOct 1, 2025California Civil Rights Departmentcalcivilrights.ca.gov
EUAI Act — GPAI model obligationsProviders of general-purpose AI models must keep technical documentation, publish a training-data summary, and maintain a copyright policy; systemic-risk models add evaluation, incident-reporting and cybersecurity duties.GeneralIn forceSince Aug 2, 2025European Commissiondigital-strategy.ec.europa.eu
US FederalSEC AI-washing enforcementThe SEC charges firms that misrepresent AI use to investors (Delphia / Global Predictions 2024; Presto 2025; Nate criminal case 2025); FY2026 exam priorities review AI-capability claims. The 2023 predictive-data-analytics conflicts rule was withdrawn June 2025.FinanceIn forceActive since 2024DLA Piperdlapiper.com
New York CityLocal Law 144 (AEDT)Employers using automated employment-decision tools must obtain an independent bias audit within the prior 12 months, post the results, and give candidates at least 10 business days’ notice. Penalties $500–$1,500 per violation, per day.EmploymentIn forceSince Jul 2023NY State Comptrollerosc.ny.gov
ConnecticutAI Responsibility & Transparency Act (SB 5)Sweeping AI law signed May 27, 2026. From Oct 1, 2026, WARN Act layoff notices must state whether AI or technological change drove the layoffs, and use of automated employment-decision technology (AEDT) is no defense to a discrimination claim. The duty to tell applicants/employees they are interacting with an AEDT phases in Oct 1, 2027.EmploymentUpcomingOct 1, 2026Ropes & Grayropesgray.com
EUAI Act — high-risk systems (Annex III)High-risk AI (incl. hiring/HR and consumer credit scoring) must meet risk-management, bias-testing, human-oversight, transparency and registration duties. The original Aug 2, 2026 deadline was DEFERRED to Dec 2, 2027 for stand-alone Annex III systems (Aug 2, 2028 for AI embedded in regulated products) by the Digital Omnibus — no longer provisional: the Council gave final approval Jun 29, 2026 (Parliament Jun 16), entering into force on publication in the Official Journal.GeneralDelayed / amendedDec 2, 2027Hogan Lovellshoganlovells.com
ColoradoColorado AI Act (SB 24-205, amended by SB 26-189)The original high-risk duty-of-care framework was repealed before taking effect; SB 26-189 (May 2026) replaced it with a narrower ADMT disclosure/transparency regime effective Jan 1, 2027. Compliance guides describing the old anti-discrimination obligations are now outdated.GeneralDelayed / amendedJan 1, 2027Huntonhunton.com
US FederalFTC Proposed Policy Statement on Suppression of Accuracy in AI Systems (Section 5)Applies the FTC Act Section 5 deception ban to companies marketing AI systems, treating undisclosed distortion of outputs for ideological or other objectives as potentially deceptive; distinguishes intentional steering from technical hallucinations. Law-firm analyses read the statement as implicitly preempting parts of Colorado's AI Act. Public comment period closes July 31, 2026.GeneralDraft guidanceComments due Jul 31, 2026Federal Registerfederalregister.gov
US FederalFDA draft guidance — AI-enabled device software functionsFirst comprehensive FDA guidance for AI-enabled medical-device software across the total product lifecycle; still draft as of mid-2026. The related Predetermined Change Control Plan guidance (pre-authorizing model updates) was finalized Dec 2024.HealthcareDraft guidanceDraft, Jan 2025FDAfda.gov
US FederalExecutive Order on state AI-law preemptionSeeks to preempt or challenge state AI laws via a DOJ AI Litigation Task Force; it cannot itself repeal state statutes. Congress rejected a state-AI moratorium twice in 2025. No state law has been preempted yet — keep complying with state law.GeneralProposed / contestedDec 11, 2025Gibson Dunngibsondunn.com

How this is compiled

  • Primary sources only. Each row links the regulator, the statute, or a law-firm analysis of it — not summaries of summaries.
  • Status over hype. We mark what is actually in force versus upcoming, delayed, contested, or still draft, so you do not build compliance work on a rule that moved or was repealed.
  • Spot a law we’re missing, or a status that changed? Tell us.

Informational only · not legal advice