The AI compliance deadline calendar

This calendar lists the main dates on which obligations in the register start to apply. The nearest are the end of Egypt's data-protection transition period and the report due from Virginia's technology commission on verifying AI models, both on November 1.

• February 2, 2025DoneEU prohibited practices and AI literacy duties apply
  The AI Act's banned-use list and workforce AI-literacy requirement took effect. European Union
• August 2, 2025DoneEU general-purpose AI obligations apply
  Transparency, copyright, and documentation duties for GPAI providers, with systemic-risk duties for the largest models. European Union
• January 1, 2026DoneCalifornia SB 53 and AB 2013; Texas TRAIGA; Illinois HB 3773; Vietnam Digital Technology Industry Law
  Frontier transparency and training-data disclosure in California, intent-based prohibitions in Texas, hiring-AI rules in Illinois, and Vietnam's digital industry statute all took effect the same day. California
• January 14, 2026DoneTaiwan's AI Basic Act in force
  Framework principles in force; a two-year clock started for aligning sectoral laws. Taiwan
• January 22, 2026DoneSouth Korea's AI Basic Act in force
  The AI Basic Act took effect, with a grace period of at least a year before most fact-finding and fines. South Korea
• January 2026DonePeru's AI regulation in force
  The implementing regulation of Law 31814 took effect ninety business days after publication; its one-to-four-year sector clocks had started on September 10, 2025. Peru
• March 1, 2026DoneVietnam's AI Law in force
  A binding comprehensive AI statute with extraterritorial reach; systems already operating have until March 1, 2027, or September 1, 2027 in healthcare, education, and finance. Vietnam
• May 19, 2026DoneUS TAKE IT DOWN Act platform removal duties apply
  Platforms hosting user content must remove reported non-consensual intimate images, AI-generated ones included, within 48 hours; the FTC enforces. United States
• June 9, 2026DoneNew York's synthetic performer ad disclosure takes effect
  Ads that use an AI-generated performer must say so, with penalties of $1,000 for a first violation and $5,000 for each later one; audio-only ads are exempt. New York
• July 1, 2026DoneTennessee SB 1580 and political deepfake disclaimer law take effect
  Tennessee's AI mental-health marketing ban and its disclaimer rule for political ads that use deepfakes of candidates took effect.
• August 2, 2026DoneEU Article 50 transparency duties and California's AI Transparency Act apply
  Chatbot disclosure and synthetic-content marking in the EU; a visible disclosure option, latent disclosures, and a public detection tool for large generative providers in California. European Union
• August 5, 2026DoneSaudi Arabia's draft AI Cybersecurity Guidelines consultation closes
  The National Cybersecurity Authority's consultation, opened July 5, closed; no final version has been published. Saudi Arabia
• August 17, 2026DoneSingapore's offense for AI-generated intimate images takes effect
  Producing intimate images without consent, including wholly AI-generated ones, becomes a crime punishable by up to two years in prison. Singapore
• August 28, 2026DoneVietnam's national AI portal opens
  The Ministry of Science and Technology put the national AI portal and database, ai.dda.gov.vn, into operation for sandbox registration, risk-classification notifications, system registration, and incident reports, eight weeks later than first planned. Vietnam
• September 10, 2026DonePeru: first sector wave reaches its compliance deadline
  Health, education, justice, security, economy, and finance, along with the branches of government and the constitutionally autonomous bodies, reach the end of their one-year period under the implementing regulation. Peru
• September 30, 2026DoneCalifornia's governor signs the session's AI bills and vetoes three
  All tracked bills signed: SB 1000 (Chapter 861, Transparency Act threshold removed, in force immediately), SB 947 (Chapter 859, operative July 1, 2027), AB 1609 (Chapter 733, signed September 28), AB 1979 (Chapter 854), AB 2713 (Chapter 856), and AB 1331 (Chapter 851), the last four from January 1, 2027, plus AB 1883, SB 503, SB 951, SB 1111, and SB 1159. Vetoed: SB 903, AB 2575, and AB 2656. California
• October 1, 2026DoneConnecticut's CART Act takes effect: first wave
  The amendment making algorithmic tools no defense to discrimination claims, frontier-developer whistleblower protections, AI-related layoff disclosures, and provenance rules, plus a narrower subscription rule from Public Act 26-100. Connecticut
• October 9, 2026DoneSubmissions close on Australia's AI standards consultation
  The Department of the Prime Minister and Cabinet's paper proposes tiered mandatory standards for data centers from 30 MW, heavier at 100 MW or a 150 MW portfolio, and safety, security, and incident-reporting duties for frontier AI developers, ahead of AI standards legislation. Australia
• October 17, 2026DoneComments close on China's draft regulation on minors online
  The draft State Council regulation would bar virtual-companion services for minors on every network service and route AI services that could affect minors' cognition through minors mode for users under 16. China
• October 26, 2026DoneColorado's rulemaking hearing and final comments on the ADMT and chatbot safety rules
  The attorney general set the rulemaking hearing for 10 am in Denver on the interim draft posted October 6, with written comments due by 11:59 pm Mountain time that day, or by the last day of the hearing if it runs longer. The rules under SB 26-189 and HB 26-1263 must be in place for the January 1, 2027 effective date. Colorado
• November 1, 2026AheadEgypt's data-protection transition period ends
  The one-year transition under the executive regulations to Law 151 of 2020 (Decree 816/2025) closes; law-firm analyses put the date at October 31 or November 1, 2026. Egypt
• November 1, 2026AheadVirginia: report due on independent verification of AI models
  The Joint Commission on Technology and Science must report on a framework for independent organizations that would verify AI models and applications against safety standards, under Chapters 425 and 426 of the 2026 Acts of Assembly. Virginia
• November 2, 2026AheadChina's legislature takes up ratification of the World AI Cooperation Organization agreement
  The NPC Standing Committee meets from November 2 to 6; its recommended agenda includes the State Council's request to ratify the agreement establishing the World AI Cooperation Organization. China
• November 16, 2026AheadExpert recommendations due under California's Executive Order N-9-26
  The order sets November 16 for the Government Operations Agency's recommendations, developed with the Office of Emergency Services and four experts named on September 23 (Jason Goldman, Gillian Hadfield, Alondra Nelson, and Rob Reich), on onsite verification at frontier labs, an emergency shutoff for frontier models, and incident reporting that includes loss of control. California
• November 18, 2026AheadNinth Circuit argument in x.AI v. Bonta, the challenge to California's AB 2013
  xAI appeals the March 4, 2026 denial of a preliminary injunction against California's training-data transparency law, which it challenges on First Amendment and trade-secret grounds. California
• November 20, 2026AheadThe UK Information Commission's call for evidence on agentic AI closes
  Six weeks of evidence from developers, deployers, and experts on the data protection risks of agentic AI, feeding the regulator's forthcoming statutory code of practice on AI and automated decision-making. United Kingdom
• November 28, 2026AheadProposed federal definition of 'super intelligence' due under Executive Order 14434
  The science adviser's 60-day deadline to submit to the President proposed legislative language for a federal definition and whether it should supersede the statutory definition of AI; the order does not require publication. United States
• November 29, 2026AheadSpain votes in a general election after its AI bills lapsed
  The Cortes were dissolved on October 6, 2026, which ended the AI governance bill, the honor and image bill, and the minors bill. The new chambers meet on December 23; a new government would have to re-table the texts. Spain
• November 2026AheadNew York opens registration for large frontier developers under the RAISE Act
  The governor's September 21 announcement: large frontier developers register with the state from November 2026, ahead of compliance from January 2027, overseen by the DIGIT office in the Department of Financial Services. New York
• November 30, 2026AheadAustralia's parliamentary AI committee delivers its final report
  The Joint Select Committee on Artificial Intelligence, which wrote to the Prime Minister on October 12 recommending incident reporting by frontier labs and embedded independent evaluators, must report by this date. Australia
• December 2, 2026AheadEU ban on non-consensual intimate imagery generators; watermark grace period ends
  The Digital Omnibus's Article 5 addition takes effect and the transitional accommodation on marking closes. European Union
• December 7, 2026AheadComments close on Treasury's proposed ban on US investment in super intelligence development in countries of concern
  The proposed rules under the COINS Act would make the development of any SI system by a covered foreign person in a country of concern a prohibited technology for US investors, with no technical threshold. United States
• December 10, 2026AheadAustralia's Privacy Act transparency duties for automated decisions apply
  Organizations covered by the Privacy Act must describe in their privacy policies the kinds of personal information used in automated decisions that could significantly affect individuals' rights or interests, and the kinds of decisions made. Australia
• End of 2026AheadSwitzerland's AI consultation bill due
  The FDJP's draft implementing the Council of Europe convention, covering transparency, data protection, non-discrimination, and oversight. Switzerland
• January 1, 2027AheadColorado's ADMT and chatbot safety laws, New York's RAISE Act, Illinois' AI Safety Measures Act, companion-chatbot rules in Connecticut and Washington, California's SB 1050 and 2026 package, and Utah's AI provenance law take effect
  Colorado's replacement statute, New York's frontier transparency law, and Illinois' frontier safety act all begin. Companion-chatbot rules start in Colorado, Connecticut, and Washington, California ads that use a synthetic performer must disclose it, and California bans companion chatbots in toys. Utah requires large generative AI providers to embed provenance data in AI images, video, and audio. California's AB 1609, AB 1979, SB 503, AB 1331, AB 1883, AB 2713, SB 951, SB 1111, and SB 1159 also start, and businesses that use automated decisionmaking technology for significant decisions must meet California's CCPA rules on notice, opt-out, and access. Colorado Utah New York Illinois Connecticut Washington California
• February 1, 2027AheadWashington's AI content disclosure law (HB 1170) takes effect
  Users must be told when content is created or modified by AI; the attorney general enforces. Washington
• March 1, 2027AheadVietnam: compliance deadline for AI systems already operating
  Systems in operation before the AI Law took effect must comply by this date, except in healthcare, education, and finance. Vietnam
• April 22, 2027AheadPeru: national public entities must have designated an AI Officer
  Directive 002-2026-PCM/SGTD gives ministries, constitutionally autonomous bodies, the judiciary, Congress, state companies, and public universities six months from its entry into force on October 22, 2026; regional governments get nine months, and Type A and C municipalities and local public companies get twelve. Peru
• July 1, 2027AheadCalifornia's Adam's Law duties and SB 947 begin; Connecticut's surveillance-pricing ban takes effect
  Companion-chatbot operators in California must meet SB 1119's safety duties, with first independent audits due January 1, 2029. Connecticut bars retailers and delivery services from pricing with tracked personal data. California Connecticut
• August 2, 2027AheadEU member states' national AI regulatory sandboxes must be operational
  As amended by the Digital Omnibus, Regulation (EU) 2026/1744, Article 57 requires each member state to have at least one national AI regulatory sandbox operational by this date. European Union
• September 1, 2027AheadVietnam: compliance deadline for healthcare, education, and finance AI systems
  The longer transition for systems already operating in these three sectors ends. Vietnam
• September 10, 2027AheadPeru: second sector wave reaches its compliance deadline
  Private uses in transport, commerce, and labor, small enterprises, and a second group of public entities reach the end of their two-year period under the implementing regulation. Peru
• October 1, 2027AheadConnecticut employment-tool developer and deployer duties apply
  For automated employment decision tools deployed from this date, developers must give deployers the information they need, and deployers must disclose the tools and give written notice before decisions. Connecticut's verification-organization pilot also begins. Connecticut
• December 2, 2027AheadEU high-risk (Annex III) obligations apply
  Recruitment, credit, education, law enforcement, and other consequential-decision systems face the full high-risk regime. European Union
• January 1, 2028AheadIllinois independent audit and framework duties begin; Connecticut's platform duties for minors apply
  Large frontier developers must publish their frontier AI framework and retain an independent third party for the first annual compliance audit under the Safety Measures Act. Connecticut's online safety duties for platforms used by minors take effect. Illinois Connecticut
• January 2028AheadTaiwan's two-year legal alignment window closes
  Agencies must have aligned or amended sectoral laws under the AI Basic Act. Taiwan
• August 2, 2028AheadEU product-embedded (Annex I) obligations apply
  AI inside regulated products such as medical devices and machinery reaches full compliance. European Union