Skip to main content

AI Regulation Deadline Calendar: 2026 and 2027 Dates

Only dates we could tie to an official text made the list; the rest came off, and the correction at the end says which.

Checked against official sources on . How we check

VerdictJanuary 1, 2027 is the date to plan around: California's ADMT rules and SB 951, Colorado's rewritten AI law and Illinois SB 315 all start that day. Before that come the FCC's TCPA opt-out rule on November 25, two EU AI Act duties on December 2, the EU Product Liability Directive on December 9 and Australia's APP 1.7 privacy policy duty on December 10.

9 min readBy Johnie T YoungUpdated 14 days ago
AI Regulation Deadline Calendar: 2026 and 2027 Dates
Image: Unsplash.

Sixteen of the twenty-five upcoming dates below fall between December and mid-January. Every row links the official text we read and the day we read it. If we couldn't find an official source for a date, it isn't here.

Upcoming deadlines

Date Where What starts or closes Who has to act Official source (checked)
Nov 25, 2026 US federal FCC's revised TCPA consent-revocation rule takes effect: a caller may name one exclusive way to opt out, and may treat an opt-out from one category of informational calls as covering only that category (our guide) Businesses that make informational robocalls or robotexts; AI-generated voices have counted as artificial voices since the FCC's February 2024 ruling (FCC 24-17) Federal Register, FR Doc. 2026-21821, Oct 31, 2026
Nov 27, 2026 India Comments close on TRAI's draft Guidelines on Responsible Use of AI in Telecommunication Networks: board-approved AI policy, an inventory of every AI system, high/medium/low risk classes, critical and high-severity AI incidents reported to TRAI, and supplier disclosures and audit terms in vendor contracts (our summary) TRAI invites written comments from stakeholders, by email to [email protected]; licensed telecom operators and their AI suppliers have the most at stake TRAI draft guidelines; TRAI Press Release No. 145/2026, Nov 2, 2026
Nov 28, 2026 US federal Executive Order 14434 (the "Super Intelligence" order of September 29) gives the Assistant to the President for Science and Technology 60 days to submit proposed legislative language for a federal definition of "Super Intelligence" to the President. It is a proposal to the President, not a rule, and nothing requires it to be published No one outside the federal government; watch it because a new statutory definition could later change which systems an AI law covers EO 14434, FR Doc. 2026-20321, Sec. 3(b), read in the govinfo.gov copy, Nov 2, 2026
Dec 2, 2026 EU Generators placed on the market before Aug 2, 2026 must meet Article 50(2): outputs marked as AI-generated in a machine-readable format (new Article 111(4)) Providers of AI systems, including general-purpose ones, that generate audio, images, video or text Regulation (EU) 2026/1744, Oct 15, 2026
Dec 2, 2026 EU New bans apply: AI that generates intimate images of an identifiable person without explicit consent, or child sexual abuse material (Article 5(1)(ba) and (bb)) Providers, and deployers who use a system for that purpose (Article 5(1a)) Regulation (EU) 2026/1744, Oct 15, 2026
Dec 9, 2026 EU New Product Liability Directive applies to products placed on the market or put into service after this date; "product" includes software Manufacturers, including software and AI providers; free and open-source software outside a commercial activity is excluded Directive (EU) 2024/2853, Oct 15, 2026
Dec 10, 2026 Australia APP 1.7 to 1.9 apply: if a computer program makes, or does something substantially and directly related to making, a decision that could significantly affect someone's rights or interests using their personal information, the privacy policy must name the kinds of personal information and decisions involved (our template) APP entities under the Privacy Act 1988; the OAIC counts rule-based programs and generative AI, including chatbots Privacy and Other Legislation Amendment Act 2024, Sch. 1, Part 15; OAIC fact sheet, Oct 27, 2026
Dec 18, 2026 New York City Comments close on DCWP's proposed rules against deceptive practices in automated customer service, the chatbot and phone menu "doom loop" rule; the rules page lists the public hearing that day at 11:00 AM. DCWP's notice PDF still says November 19 for both, so check the rules page before you plan to speak (our checklist) Anyone can comment; businesses whose chatbots or phone menus serve New York City consumers have the most at stake NYC Rules; DCWP notice, Oct 31, 2026
Dec 28, 2026 US federal HHS's HTI-5 rule takes effect: certified health IT no longer has to support source attributes (the "model card" fields) or intervention risk management for decision support interventions (§ 170.315(b)(11)(iv) to (vi)); three other amendments wait until Jan 1, 2027. Our HTI-5 vendor questions list what to ask for before then Developers of certified health IT, and the providers who buy it Federal Register, FR Doc. 2026-21943, Oct 31, 2026
Jan 1, 2027 Colorado SB 26-189 takes effect; it repeals and reenacts the 2024 Colorado AI Act (SB 24-205) Developers of ADMT used to materially influence a consequential decision, and deployers who use it Colorado Attorney General, Oct 15, 2026
Jan 1, 2027 Colorado HB 26-1263, the Chatbot Safety Act: age estimation, AI disclosure, safeguards for teen users Chatbot operators Colorado Attorney General, Oct 15, 2026
Jan 1, 2027 California CCPA automated decisionmaking rules: pre-use notice (§ 7220), opt-out (§ 7221), access (§ 7222); a business already using ADMT for a significant decision must comply by this date (§ 7200(b)) Businesses under the CCPA that use ADMT for significant decisions about consumers CPPA approved regulations text, Oct 15, 2026
Jan 1, 2027 California SB 951 (Chapter 860): a Cal-WARN notice for a layoff caused in whole or in substantial part by AI must give the number, occupations and work location of affected workers, the job functions being automated and the type of technology, under the heading "This notice is for a technology displacement" Employers with a covered establishment (75 or more people in the past 12 months) SB 951 chaptered text, Oct 15, 2026
Jan 1, 2027 California SB 690 (Chapter 976): claims under Penal Code § 638.51 over a website or app can be brought only by the Attorney General, including pending claims filed within two years before that date Website and app operators facing pen register suits SB 690 chaptered text, Oct 15, 2026
Jan 1, 2027 Connecticut SB 5 (Public Act 26-15), Sec. 5: AI companions need a protocol that detects expressions of suicide, self-harm or violence, and must tell users they're talking to an AI (at least hourly for minors) Operators of AI companions Public Act 26-15, Oct 15, 2026
Jan 1, 2027 Illinois SB 315, the Artificial Intelligence Safety Measures Act, takes effect The largest AI developers Governor's release, July 6, 2026, Oct 15, 2026
Jan 1, 2027 New York RAISE Act (Chapter 699 of 2025, final text in Chapter 96 of 2026) takes effect: a transparency report with each new frontier model, critical safety incidents reported within 72 hours to a new office in the Department of Financial Services; large frontier developers also publish a frontier AI framework Frontier developers (models trained with more than 10²⁶ operations); the framework duty only for those with over $500 million in annual revenue New York State Assembly, A9449 (same as S8828), Oct 16, 2026
Jan 1, 2027 New York S.9051B/A.10379C (Chapter 368 of 2026), General Business Law Article 48: an operator may not offer the listed "unsafe AI companion features" to a New York user unless it has used approved methods to determine the user isn't a minor; the attorney general can seek up to $25,000 per violation (our guide) Chatbot operators with New York users; a customer service bot is exempt only if it can't respond on topics outside that purpose New York State Assembly, A10379, Oct 31, 2026
Jan 14, 2027 UK Law Commission consultation on product liability closes; the proposals would bring software, AI systems and online platform design into "product" Anyone; software and AI producers most directly Law Commission, Oct 15, 2026
Jul 1, 2027 California SB 1119 (Chapter 190), Adam's Law: before releasing a new or substantially modified companion chatbot, the operator must perform and document a risk assessment for child users; the ban on showing children cross-context behavioral ads starts the same day Operators of companion chatbots available in California SB 1119 chaptered text, Oct 16, 2026
Aug 2, 2027 EU General-purpose AI models placed on the market before Aug 2, 2025 must comply with the AI Act (Article 111(3)); models launched since then have been covered from launch (GPAI checklist) Providers of those general-purpose AI models AI Act Service Desk, Article 111, Oct 31, 2026
Oct 1, 2027 Connecticut SB 5, Secs. 8 and 9: deployers must disclose automated employment decision tools to employees and applicants; developers must give deployers what they need Deployers and developers of tools deployed on or after this date Public Act 26-15, Oct 15, 2026
Oct 7, 2027 Singapore MAS Guidelines on AI risk management take effect (Sections 3 and 4; Sections 5 and 6 a year later) Financial institutions supervised by MAS MAS media release, Oct 15, 2026
Dec 2, 2027 EU High-risk duties (Chapter III, Sections 1 to 3) for systems listed in Annex III, such as recruitment, education and credit scoring. For the provider's quality management system (Article 17), EN 18286 has been a harmonised standard since Oct 14, 2026 (what it covers) Providers and deployers of Annex III systems Regulation (EU) 2026/1744, Oct 15, 2026
Aug 2, 2028 EU High-risk duties for AI in products covered by Annex I Providers of AI that is a safety component of regulated products Regulation (EU) 2026/1744, Oct 15, 2026

California's two bills carry no urgency clause, so they take effect on the January 1 after a 90-day period from enactment (California Constitution, art. IV, § 8(c)). Both were signed on September 30.

Colorado Attorney General rulemaking page stating that written comments are accepted through October 26, 2026 at 11:59 p.m. MST

Screenshot: ADMT and Chatbot Safety rulemaking page, Colorado Attorney General, checked October 15, 2026 (highlight added).

Colorado's written comments on the attorney general's draft ADMT and Chatbot Safety rules closed on October 26, so that row came off the list. Comments are posted publicly, and our comment letter template is still there if you want to see what we suggested raising. The UK consultation is now the one row that still lets you shape the rule rather than just follow it.

Already in force

Date Where What applies Who it covers Official source (checked)
Feb 2, 2025 EU AI Act Chapters I and II, including the prohibited practices in Article 5 Providers and deployers in the EU market Regulation (EU) 2024/1689, Art. 113, Oct 15, 2026
Aug 2, 2025 EU Chapter V obligations for general-purpose AI models, plus governance and penalties (except Article 101) General-purpose AI model providers Regulation (EU) 2024/1689, Art. 113, Oct 15, 2026
Jan 1, 2026 Texas HB 149, the Texas Responsible Artificial Intelligence Governance Act Use of AI systems in Texas Texas Legislature Online, Oct 15, 2026
Jul 27, 2026 EU Digital Omnibus on AI, Regulation (EU) 2026/1744, in force (published July 24, in force on the third day) Everyone covered by the AI Act Regulation (EU) 2026/1744, Oct 15, 2026
Aug 2, 2026 EU AI Act general application, including Article 50 transparency duties for systems launched from this date; Commission fines on model providers (Article 101) Providers and deployers; general-purpose AI model providers Regulation (EU) 2024/1689, Art. 113, Oct 15, 2026
Sep 30, 2026 California SB 1000 (Chapter 861), an urgency statute in effect on signing: amends the AI Transparency Act and drops the user threshold from "covered provider" Covered providers of generative AI systems SB 1000 chaptered text, Oct 15, 2026
Oct 1, 2026 Connecticut SB 5: key-terms notice for AI subscriptions (Sec. 1), frontier developer whistleblower protections (Sec. 2), an AEDT is no defense to a discrimination complaint (Sec. 13), provenance data from generators with over a million monthly users (Sec. 15), WARN filers say whether layoffs relate to AI (Sec. 26) Subscription AI providers, frontier developers, employers, large generators Public Act 26-15, Oct 15, 2026
Oct 1, 2026 Maryland HB 895 (Chapter 154), the Protection From Predatory Pricing Act: no dynamic pricing and no pricing from consumer personal data Food retailers and third-party delivery services Maryland General Assembly, Oct 15, 2026

EUR-Lex text of Regulation (EU) 2026/1744 adding Article 111(4) and amending Article 113 of the AI Act with the December 2, 2026 dates

Screenshot: Regulation (EU) 2026/1744, Article 1, points (39) and (40), EUR-Lex, checked October 15, 2026 (highlight added).

US federal: what happened to the EO 14365 deadlines

Executive Order 14365, signed December 11, 2025, set the federal deadlines most often quoted as "the March 2026 AI deadlines." We read the order's text and then looked for each promised document at its official source on November 2, 2026. Only one of the three March 11 tasks has produced anything we could find, and that one is still a proposal.

EO 14365 section What it ordered Due What we found (checked Nov 2, 2026) Official source
Sec. 3 Attorney General sets up an AI Litigation Task Force to challenge state AI laws Jan 10, 2026 (30 days) Done: Attorney General memorandum of January 9, 2026 DOJ memorandum
Sec. 4 Commerce publishes an evaluation of existing state AI laws that names "onerous" laws and laws to refer to the Task Force Mar 11, 2026 (90 days) Not in the Federal Register: a search of Commerce Department documents mentioning "state AI laws" published since January 1, 2026 returns none Federal Register
Sec. 5(a) Commerce, through NTIA, issues a BEAD Policy Notice making states with those laws ineligible for non-deployment funds Mar 11, 2026 Not issued as far as NTIA's own FAQ shows: BEAD FAQs Version 25 (October 2026) says approval for all non-deployment activities is rescinded and "Further guidance will be provided" NTIA BEAD FAQs v25, Q2.21 and Q2.30
Sec. 6 FCC opens a proceeding on a federal AI reporting and disclosure standard 90 days after the Sec. 4 evaluation is published Not triggered, since we found no Sec. 4 evaluation EO 14365
Sec. 7 FTC issues a policy statement on when state laws requiring changes to truthful AI outputs are preempted Mar 11, 2026 Proposed only: "Policy Statement Concerning the Suppression of Accuracy in Artificial Intelligence Systems," published as a notice on July 7, 2026; no final version in the Federal Register FR Doc. 2026-13628

None of this switched off a state law. Section 9(c) of the order says it does not "create any right or benefit, substantive or procedural, enforceable at law," so a state AI law stays in force until a court or Congress says otherwise. The Colorado, California, New York and Illinois rows above still apply on their dates. Our federal preemption guide covers what the Task Force could challenge, and the AI law tracker keeps the status of each order.

Penalties we confirmed

Law Maximum penalty Who enforces Official source (checked)
EU AI Act, prohibited practices EUR 35 million or 7% of worldwide annual turnover, whichever is higher (Article 99(3)) Member States set the enforcement rules Regulation (EU) 2024/1689, Oct 15, 2026
EU AI Act, general-purpose AI models 3% of worldwide annual turnover or EUR 15 million, whichever is higher (Article 101) European Commission Regulation (EU) 2024/1689, Oct 15, 2026
California AI Transparency Act, as amended by SB 1000 $5,000 per violation (§ 22757.4) Attorney General, a city attorney or a county counsel SB 1000 chaptered text, Oct 15, 2026
Connecticut SB 5 Up to $1,000 per violation of the frontier whistleblower rules; for hiring-tool violations through December 31, 2027, the AG first sends a notice with a chance to cure where a cure is possible; no private right of action Connecticut Attorney General Public Act 26-15, Oct 15, 2026

What we could not confirm: ilga.gov didn't load for us on October 15, so the Illinois row rests on the Governor's release, not the statute; we left out its duties for that reason. The Connecticut legislature's site timed out too, so we read Public Act 26-15 from the Internet Archive copy of the official PDF captured October 7, 2026. The New York Senate and Governor's sites blocked our browser on October 16, so the New York row rests on the bill text and history on the Assembly's site, which shows the amendment signed as Chapter 96 on March 27, 2026 and moving the start date to January 1, 2027. The old federal rows (an FTC rulemaking, a second Operation AI Comply) had no official source behind their dates, so they're gone.

What changed since the last review

Reviewed October 15, 2026, two days after the previous pass. New rows: the two EU AI Act dates on December 2, the Product Liability Directive on December 9, Colorado's comment deadline and its Chatbot Safety Act, the CCPA automated decisionmaking rules, SB 951, SB 690, SB 1000, Illinois SB 315, the UK consultation, the MAS guidelines and Maryland HB 895. On October 16 we added New York's RAISE Act and California's SB 1119 (Adam's Law) after reading both enacted texts. On October 27 we added Australia's APP 1.7 row after reading the commencement table and Part 15 of the amending Act, and took off Colorado's comment deadline, which passed on October 26.

On October 31 we added five rows, each read at its official source that day: the FCC's TCPA opt-out rule (November 25), New York City's comment deadline and hearing (December 18), HHS's HTI-5 rule (December 28), New York's chatbot law S.9051B (January 1, 2027) and the August 2, 2027 deadline for general-purpose AI models already on the market before August 2, 2025. The HTI-5 row is the odd one out: it removes an AI transparency duty rather than adding one. The Federal Register site turned away our scripted download, so we read the HTI-5 text in the govinfo.gov copy of the same October 28 issue.

On November 2 we added two rows and one section. The rows are TRAI's comment deadline on its draft telecom AI guidelines (November 27), read in TRAI's draft and press release, and the 60-day task in Executive Order 14434 (November 28), read in the govinfo.gov copy of the order. The new section on EO 14365 lists each March 2026 federal deadline next to what we found at the official source: a Federal Register search, NTIA's BEAD FAQs Version 25 and the FTC's July 7 notice on November 2, and the DOJ memorandum, which we read on October 31 for our AI law tracker. DOJ's site refused our automated access on November 2, so that one row was not re-opened.

We took out every row we couldn't trace to an official text, including the Colorado SB 205 dates, which SB 26-189 replaced before they took effect. The full list is in the correction below. There's no embedded post on this page: no single thread or official post covers the whole calendar, so the screenshots show the primary texts instead.

California SB 951 chaptered text showing Labor Code section 1401(d) and the statement This notice is for a technology displacement

Screenshot: SB 951, Labor Code § 1401(d) as amended, California Legislative Information, checked October 15, 2026 (highlight added).

How to use this calendar

Give each row one owner and a calendar entry two weeks before the date. January 1 lands on three teams at once in California: HR (SB 951 notices), privacy (the ADMT rules) and the web team (SB 690). Check which of them you are before December.

If you ship generative features into the EU, test your output marking before December 2; a system launched after August 2, 2026 has needed it since launch. For the documents behind each row, start with the Connecticut SB 5 checklist, the California ADMT templates and the Product Liability Directive guide.

Correction (October 13, 2026): Earlier versions listed Connecticut SB 5 as requiring opt-out, annual bias audits, four-year records and AEDT notices from October 1, 2026, called it "SB 2" and "enacted 2025", and said penalties were left to rulemaking. Public Act 26-15 (approved May 27, 2026) has none of those; the notices start October 1, 2027. We also corrected the EU AI Act GPAI date (August 2, 2025) and moved the Annex I high-risk date to August 2, 2028 under Regulation (EU) 2026/1744.

Correction (October 15, 2026): We rebuilt this calendar from official sources and removed rows we couldn't confirm: Colorado SB 205 taking effect February 1, 2026 (SB 26-189 repealed and reenacted it; duties start January 1, 2027), an "Illinois AEDT annual audit" in 2025 and 2027, "CalAI amendments (SB 1047 provisions)", a second wave of the FTC's Operation AI Comply, a "Washington AI Likeness Protection Act", a "Texas AI Liability Act", a 2027 FTC AI rulemaking, Connecticut SB 4, a sector table of ongoing federal obligations, and penalty rows for Colorado SB 205, NYC Local Law 144 and FTC Section 5. Earlier versions also listed August 2, 2026 as the start of general-purpose AI model obligations (they applied from August 2, 2025), cited an unsourced "EU AI Office confirmed, May 2026" line, and described Maryland's pricing law as awaiting rulemaking; HB 895 took effect October 1, 2026.

Legal disclaimer

This article is published for informational and educational purposes only. It does not constitute legal, regulatory, or professional compliance advice and should not be relied upon as such. AI governance requirements vary by jurisdiction, industry, and organizational context. Always consult a qualified legal or compliance professional before implementing policies or making decisions with regulatory implications.

About the author

Johnie T Young

Governance & Regulation Desk

Covers AI laws, regulator orders and court rulings, and what each one means for a team of ten. Posts start from the statute or the order itself, with every fact traced to a row in a source log.

  • Starts from the statute or order
  • Every fact traced to a source row
  • Dated checks on each post