BEGIN:VCALENDAR VERSION:2.0 PRODID:-//AI News Today//AI Compliance Calendar//EN CALSCALE:GREGORIAN METHOD:PUBLISH X-WR-CALNAME:AI Compliance Deadlines — AI News Today X-WR-CALDESC:Dated AI compliance obligations, each verified against its pr imary source. X-PUBLISHED-TTL:PT12H REFRESH-INTERVAL;VALUE=DURATION:PT12H BEGIN:VEVENT UID:vn|prime-minister-s-decision-no-33-2026-qd-ttg-of-30-june-2026-issuing -th|high-risk-ai-list-takes-effect-46-named-ai-systems-across-6|2026-08-1 5@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260815 DTEND;VALUE=DATE:20260816 SUMMARY:Vietnam: High-risk AI list takes effect: 46 named AI systems acros s 6 sectors become 'high-risk' under the Law on AI 134/2025/QH15\, trigge ring the strict-management duties (operating principles and compliance ro admap set by the Decision) DESCRIPTION:High-risk AI list takes effect: 46 named AI systems across 6 s ectors become 'high-risk' under the Law on AI 134/2025/QH15\, triggering the strict-management duties (operating principles and compliance roadmap set by the Decision)\nPrime Minister's Decision No. 33/2026/QD-TTg of 30 June 2026 issuing the List of high-risk artificial intelligence systems (Quyết định 33/2026/QĐ-TTg — Danh mục hệ thống trí tuệ n hân tạo có rủi ro cao)\nSigned 30 June 2026 by Deputy PM Hồ Quố c Dũng. Sectors: education\; ethnic and religious affairs\; health\; ban king\; legal proceedings (tố tụng)\; transport (transport alone accou nts for 31 of the 46 systems). The Decision sets both the operating princ iples and the compliance roadmap for systems on the list.\nApplies to: pr ovider\, deployer\nWho: all\nPenalty: State AI authority may require susp ension or termination of a system it finds poses a risk of serious harm\n Source says: "Ngày có hiệu lực: 15-08-2026 // 'Effective date: 15 A ugust 2026'. Người ký: Hồ Quốc Dũng\; Ngày ban hành: 30-06-202 6."\nSource: https://vanban.chinhphu.vn/?pageid=27160&docid=218658&classi d=1&typegroupid=5\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-complia nce-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Law applies TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Vietnam: High-risk AI list takes effect: 46 named AI systems a cross 6 sectors become 'high-risk' under the Law on AI 134/2025/QH15\, tr iggering the strict-management duties (operating principles and complianc e roadmap set by the Decision) END:VALARM END:VEVENT BEGIN:VEVENT UID:nz|biometric-processing-privacy-code-2025-end-of-the-transition-period -fo|organisations-that-were-already-carrying-out-automated-biome|2026-08- 03@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260803 DTEND;VALUE=DATE:20260804 SUMMARY:New Zealand: Organisations that were ALREADY carrying out automate d biometric processing before 3 November 2025 must\, from this date\, com ply in full with the Code's 13 rules — including the Rule 1 proportiona lity assessment\, notification duties and the limits on biometric categor isation DESCRIPTION:Organisations that were ALREADY carrying out automated biometr ic processing before 3 November 2025 must\, from this date\, comply in fu ll with the Code's 13 rules — including the Rule 1 proportionality asse ssment\, notification duties and the limits on biometric categorisation\n Biometric Processing Privacy Code 2025 — end of the transition period f or existing biometric processing\nThe Privacy Commissioner gave a nine-mo nth grace period from commencement for legacy biometric deployments (faci al recognition in retail and security\, voice authentication\, age-estima tion systems). That grace period has now expired\, so pre-existing system s are fully in scope.\nApplies to: deployer\nWho: agencies already carryi ng out biometric processing before 3 November 2025\nPenalty: Interference with privacy under the Privacy Act 2020: complaints\, compliance notices \, and Human Rights Review Tribunal proceedings including damages\nSource says: "The BPPC came into force on 3 November 2025\, but agencies alread y using biometrics had a nine-month grace period to move to the new set o f rules. That transition period ended on 3 August 2026."\nSource: https:/ /www.privacy.org.nz/privacy-principles/codes-of-practice/biometric-proces sing-privacy-code/\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compli ance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:New Zealand: Organisations that were ALREADY carrying out auto mated biometric processing before 3 November 2025 must\, from this date\, comply in full with the Code's 13 rules — including the Rule 1 proport ionality assessment\, notification duties and the limits on biometric cat egorisation END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689|transparency-obligations-for-ai-i nteracting-with-people-synt|2026-08-02@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260802 DTEND;VALUE=DATE:20260803 SUMMARY:European Union: Transparency obligations for AI interacting with p eople\, synthetic content marking and deepfake disclosure apply (Art. 50) DESCRIPTION:Transparency obligations for AI interacting with people\, synt hetic content marking and deepfake disclosure apply (Art. 50)\nEU AI Act (Regulation (EU) 2024/1689)\nChatbots must disclose they are machines\; s ynthetic audio\, image\, video and text must be machine-readably marked\; deepfakes and AI-generated public-interest text must be labelled. Also t he date the Act's penalties regime and Member State sandboxes become oper ational.\nApplies to: provider\, deployer\nPenalty: Up to EUR 15 000 000 or 3% of worldwide annual turnover (Art. 99(4))\nSource says: "It shall a pply from 2 August 2026."\nSource: https://ai-act-service-desk.ec.europa. eu/en/ai-act/article-113\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai- compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Transparency obligations for AI interacting wi th people\, synthetic content marking and deepfake disclosure apply (Art. 50) END:VALARM END:VEVENT BEGIN:VEVENT UID:cn|interim-measures-for-the-administration-of-anthropomorphic-ai-inter act|anthropomorphic-companion-emotional-ai-interaction-service-p|2026-07- 15@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260715 DTEND;VALUE=DATE:20260716 SUMMARY:China (People's Republic of China): Anthropomorphic (companion/emo tional) AI interaction service providers must comply in full: AI-disclosu re and content labelling\, 2-hour continuous-use reminders\, minors prote ctions\, security assessment filed with the provincial cyberspace adminis tration\, and algorithm filing DESCRIPTION:Anthropomorphic (companion/emotional) AI interaction service p roviders must comply in full: AI-disclosure and content labelling\, 2-hou r continuous-use reminders\, minors protections\, security assessment fil ed with the provincial cyberspace administration\, and algorithm filing\n 人工智能拟人化互动服务管理暂行办法 — Interim Measures for the Administration of Anthropomorphic AI Interaction Services (CAC/ND RC/MIIT/MPS/SAMR Order No. 21\, promulgated 10 April 2026)\nChina's first national-level rules aimed specifically at AI companionship/emotional-in teraction services. From this date providers must: (Art. 18) discharge th e AI-generated-synthetic-content labelling duty and take effective measur es to alert users that they are interacting with an AI service and not a natural person\, and remind users of elapsed time by dialogue or pop-up e very time continuous use exceeds 2 hours\; (Art. 14) NOT provide virtual- kin\, virtual-partner or other virtual intimate-relationship services to minors at all\, and obtain parent/guardian consent before providing other anthropomorphic interaction services to under-14s\; (Art. 22) carry out a security assessment and submit the assessment report to the provincial cyberspace administration before launching such a service or adding anthr opomorphic-interaction functionality\; (Art. 26) complete algorithm filin g\, and filing changes/cancellations\, under the Internet Information Ser vice Algorithmic Recommendation Management Provisions.\nApplies to: provi der (拟人化互动服务提供者) — service provider offering anthro pomorphic AI interaction services to the public within the PRC\nPenalty: Art. 30: handling/punishment by the cyberspace\, development-and-reform\, industry-and-information-technology and public-security departments unde r applicable laws and administrative regulations\; where no law or regula tion provides\, those departments may issue a warning\, circulate critici sm\, order correction within a time limit\, and may require measures such as suspending new user account registration or other related services. F or refusal to correct or serious circumstances: order to stop providing t he relevant service\, plus a fine of RMB 10\,000–100\,000\; where citiz ens' life and health safety are endangered with harmful consequences\, a fine of RMB 100\,000–200\,000.\nSource says: "第三十二条 本办法 自2026年7月15日起施行。"\nSource: https://www.cac.gov.cn/2026-04/ 10/c_1777558395078289.htm\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai -compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Law applies TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:China (People's Republic of China): Anthropomorphic (companion /emotional) AI interaction service providers must comply in full: AI-disc losure and content labelling\, 2-hour continuous-use reminders\, minors p rotections\, security assessment filed with the provincial cyberspace adm inistration\, and algorithm filing END:VALARM END:VEVENT BEGIN:VEVENT UID:sg|online-safety-relief-and-accountability-act-2025-bill-18-2025-passe d-5|online-safety-commission-starts-operating-and-the-statutory|2026-06-2 9@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260629 DTEND;VALUE=DATE:20260630 SUMMARY:Singapore: Online Safety Commission starts operating and the statu tory torts / directions regime commences for the first five harms — int imate image abuse\, image-based child abuse\, doxxing\, online harassment (including online sexual harassment) and online stalking. AI-generated m aterial is expressly inside the first two: an 'intimate image or recordin g' and a 'child abuse image or recording' include images or recordings 'a ltered or generated by any means'. Communicators of harmful content\, onl ine administrators\, online platforms\, internet access service providers and app distribution services must comply with Commission directions (ta kedown\, account restriction\, right of reply\, access disabling). DESCRIPTION:Online Safety Commission starts operating and the statutory to rts / directions regime commences for the first five harms — intimate i mage abuse\, image-based child abuse\, doxxing\, online harassment (inclu ding online sexual harassment) and online stalking. AI-generated material is expressly inside the first two: an 'intimate image or recording' and a 'child abuse image or recording' include images or recordings 'altered or generated by any means'. Communicators of harmful content\, online adm inistrators\, online platforms\, internet access service providers and ap p distribution services must comply with Commission directions (takedown\ , account restriction\, right of reply\, access disabling).\nOnline Safet y (Relief and Accountability) Act 2025 (Bill 18/2025\, passed 5 November 2025) — partial commencement\; Online Safety Commission begins operatio ns\nThe Act creates 13 categories of online harm\; only these five commen ce on this date\, and the remaining eight — including 'inauthentic mate rial abuse' (deepfakes) — are to be implemented progressively with NO d ate announced. Mr Francis Ng was appointed Commissioner-Designate of Onli ne Safety from 1 June 2026 to 28 June 2026 and became Commissioner on 29 June 2026.\nApplies to: platform\, online-administrator\, internet-access -service-provider\, app-distribution-service\, individual\nWho: all\nPena lty: Statutory torts giving victims civil claims\; offences for non-compl iance with Online Safety Commission directions\nSource says: "From 29 Jun e 2026\, Singaporeans will have stronger protection and faster avenues fo r relief against online harms. … Intimate image abuse\; Image-based chi ld abuse\; Doxxing\; Online harassment (including online sexual harassmen t)\; and Online stalking. … The remaining categories will be progressiv ely implemented."\nSource: https://www.mlaw.gov.sg/online-safety-commissi on-and-online-safety-relief-and-accountability-act-2025-to-start-on-29-ju ne-2026/\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calen dar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Enforcement TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Singapore: Online Safety Commission starts operating and the s tatutory torts / directions regime commences for the first five harms — intimate image abuse\, image-based child abuse\, doxxing\, online harass ment (including online sexual harassment) and online stalking. AI-generat ed material is expressly inside the first two: an 'intimate image or reco rding' and a 'child abuse image or recording' include images or recording s 'altered or generated by any means'. Communicators of harmful content\, online administrators\, online platforms\, internet access service provi ders and app distribution services must comply with Commission directions (takedown\, account restriction\, right of reply\, access disabling). END:VALARM END:VEVENT BEGIN:VEVENT UID:uk|crime-and-policing-act-2026-schedule-13-paragraph-24-amendment-of-s che|sexual-deepfake-offences-become-priority-offences-under-the|2026-06-2 9@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260629 DTEND;VALUE=DATE:20260630 SUMMARY:United Kingdom: Sexual deepfake offences become 'priority offences ' under the Online Safety Act\, extending Part 3 illegal-content duties o f user-to-user and search services to AI-generated intimate images DESCRIPTION:Sexual deepfake offences become 'priority offences' under the Online Safety Act\, extending Part 3 illegal-content duties of user-to-us er and search services to AI-generated intimate images\nCrime and Policin g Act 2026\, Schedule 13 paragraph 24 (amendment of Schedule 7 to the Onl ine Safety Act 2023)\, commenced by SI 2026/689 reg. 2(1)(z9)\nParagraph 24 of Schedule 13 adds to paragraph 28A of Schedule 7 to the Online Safet y Act 2023 the offences in s.66E (creating purported intimate image of ad ult) and s.66F (requesting the creation of purported intimate image of ad ult) of the Sexual Offences Act 2003. Listing an offence in Schedule 7 ma kes the related content 'priority illegal content'\, so Part 3 providers must take proportionate proactive measures to prevent users encountering it and to minimise the time it is present\, and must reflect it in their illegal content risk assessments.\nApplies to: providers of user-to-user services\, providers of search services\nWho: all Part 3 services in scop e of the Online Safety Act 2023\nPenalty: Ofcom enforcement under the Onl ine Safety Act 2023: fines up to the greater of GBP 18 million or 10% of qualifying worldwide revenue\, business disruption measures\, and senior manager liability for named offences\nSource says: "In Schedule 7 to the Online Safety Act 2023 (priority offences)\, in paragraph 28A (Sexual Off ences Act 2003)\, at the end insert— "(c) section 66E (creating purport ed intimate image of adult)\; (d) section 66F (requesting the creation of purported intimate image of adult).""\nSource: https://www.legislation.g ov.uk/ukpga/2026/20/schedule/13/paragraph/24\nVerified: Aug 30\, 2026\nht tps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United Kingdom: Sexual deepfake offences become 'priority offe nces' under the Online Safety Act\, extending Part 3 illegal-content duti es of user-to-user and search services to AI-generated intimate images END:VALARM END:VEVENT BEGIN:VEVENT UID:uk|crime-and-policing-act-2026-s-99-purported-intimate-image-generator s-c|offences-of-making-adapting-possessing-supplying-or-offering|2026-06- 29@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260629 DTEND;VALUE=DATE:20260630 SUMMARY:United Kingdom: Offences of making\, adapting\, possessing\, suppl ying or offering to supply a tool for creating purported (deepfake) intim ate images take effect DESCRIPTION:Offences of making\, adapting\, possessing\, supplying or offe ring to supply a tool for creating purported (deepfake) intimate images t ake effect\nCrime and Policing Act 2026\, s.99 (purported intimate image generators)\, commenced by SI 2026/689 (Commencement No.1 and Saving Prov ision) Regulations 2026\nSection 99 inserts new sections 66I-66L into the Sexual Offences Act 2003\, criminalising making or adapting 'a thing for creating\, or facilitating the creation of\, purported intimate images o f a person' and possessing\, supplying or offering to supply such a gener ator. 'Thing' is defined to include 'a program\, information in electroni c form and a service'\, which captures nudification apps\, image-generati on models and hosted services. This targets the supply side of sexual dee pfakes\, complementing the s.138 DUAA creation offence.\nApplies to: deve lopers of image-generation models\, operators of nudification services\, app stores and hosts\, individuals\nWho: all\nPenalty: Criminal offence u nder the Sexual Offences Act 2003 as amended\; see the new ss.66I-66L for the applicable penalties\nSource says: "Subject to paragraph (2)\, the f ollowing provisions of the 2026 Act come into force on 29th June 2026— ... (i) section 99 (purported intimate image generators)\;"\nSource: http s://www.legislation.gov.uk/uksi/2026/689/regulation/2/made\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Enforcement TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United Kingdom: Offences of making\, adapting\, possessing\, s upplying or offering to supply a tool for creating purported (deepfake) i ntimate images take effect END:VALARM END:VEVENT BEGIN:VEVENT UID:ca|treasury-board-directive-on-automated-decision-making-issued-under- the|legacy-automated-decision-systems-must-be-brought-into-compl|2026-06- 24@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260624 DTEND;VALUE=DATE:20260625 SUMMARY:Canada — federal: Legacy automated decision systems must be brou ght into compliance with the new/updated Directive requirements DESCRIPTION:Legacy automated decision systems must be brought into complia nce with the new/updated Directive requirements\nTreasury Board Directive on Automated Decision-Making (issued under the Policy on Service and Dig ital\; amendments dated 2025-06-24)\nFederal departments operating automa ted decision systems that were developed or procured BEFORE 2025-06-24 ha d a transition window to meet the requirements added in the 2025 amendmen t round — Algorithmic Impact Assessment completion and publication on t he Open Government Portal\, the Appendix C measures for the assigned impa ct level (notice\, explanation\, human-in-the-loop\, peer review\, monito ring\, employee training\, IT/business continuity)\, and access-to-compon ents rights for audit. Systems developed or procured after 2020-04-01 wer e already bound\; s.1.2.1 is the catch-up date for the pre-existing estat e.\nApplies to: deployer\, operator\nWho: all federal departments listed in Schedules I\, I.1 and II of the Financial Administration Act\, per the Policy on Service and Digital scope\nPenalty: No monetary penalty. Non-c ompliance is handled administratively under the Treasury Board Framework for the Management of Compliance — deputy heads must respond to non-com pliance\; consequences can include removal of delegated authority and req uiring the system be taken out of production. Section 8.3.5 requires appr oval to operate for level 4 (highest impact) systems.\nSource says: "1.2. 1 Existing automated decision systems developed or procured prior to June 24\, 2025\, will have until June 24\, 2026 to comply with the new or upd ated requirements."\nSource: https://www.tbs-sct.canada.ca/pol/doc-eng.as px?id=32592\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-ca lendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Canada — federal: Legacy automated decision systems must be brought into compliance with the new/updated Directive requirements END:VALARM END:VEVENT BEGIN:VEVENT UID:ca|treasury-board-directive-on-automated-decision-making-s-1-2-2|agent s-of-parliament-must-comply-with-the-directive-s-requir|2026-06-24@agencc y.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260624 DTEND;VALUE=DATE:20260625 SUMMARY:Canada — federal: Agents of Parliament must comply with the Dire ctive's requirements DESCRIPTION:Agents of Parliament must comply with the Directive's requirem ents\nTreasury Board Directive on Automated Decision-Making\, s.1.2.2\nSe parate transition date extending the Directive to Agents of Parliament (e .g. Office of the Auditor General\, Office of the Privacy Commissioner\, Office of the Commissioner of Official Languages\, Chief Electoral Office r). Their heads are solely responsible for monitoring and enforcing compl iance internally (s.8.3.2) and for approving level 4 systems to operate ( s.8.3.5).\nApplies to: deployer\, operator\nWho: all Agents of Parliament \nPenalty: No monetary penalty. Heads of Agents of Parliament are 'solely responsible for monitoring and ensuring compliance with this directive w ithin their organizations\, as well as for responding to cases of non-com pliance' (s.8.3.2).\nSource says: "1.2.2 Agents of Parliament will have u ntil June 24\, 2026\, to comply with the requirements."\nSource: https:// www.tbs-sct.canada.ca/pol/doc-eng.aspx?id=32592\nVerified: Aug 30\, 2026\ nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Canada — federal: Agents of Parliament must comply with the Directive's requirements END:VALARM END:VEVENT BEGIN:VEVENT UID:my|risk-mitigation-code-issued-by-the-malaysian-communications-and-mul tim|licensed-service-providers-must-label-synthetic-media-genera|2026-06- 01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260601 DTEND;VALUE=DATE:20260602 SUMMARY:Malaysia: Licensed service providers must label synthetic media: g enerated or manipulated images\, audio or video that closely resemble rea l persons\, objects\, places\, entities or events and are likely to false ly appear authentic must be clearly distinguishable through prominent lab els or markings on the service's online interface (para. 4.2.4(d)\, 'Safe Design of the Service'). The same Code also requires providers to test a nd adapt algorithmic and recommender systems against exposure to harmful content (para. 4.2.4(c)) and to run a documented harmful-content risk ass essment reviewed at least annually. DESCRIPTION:Licensed service providers must label synthetic media: generat ed or manipulated images\, audio or video that closely resemble real pers ons\, objects\, places\, entities or events and are likely to falsely app ear authentic must be clearly distinguishable through prominent labels or markings on the service's online interface (para. 4.2.4(d)\, 'Safe Desig n of the Service'). The same Code also requires providers to test and ada pt algorithmic and recommender systems against exposure to harmful conten t (para. 4.2.4(c)) and to run a documented harmful-content risk assessmen t reviewed at least annually.\nRisk Mitigation Code issued by the Malaysi an Communications and Multimedia Commission (MCMC) under s.80 of the Onli ne Safety Act 2025 [Act 866]\nThe Code specifies the measures licensed se rvice providers must implement to discharge their duty under s.13 of the ONSA 2025. It binds providers of applications services enabling communica tion between users\, and of content applications services\, that use inte rnet access service. Where synthetic content is to be identified through user or advertiser disclosures/reports\, the provider must supply accessi ble functionality and guidance for those disclosures. Providers may subst itute alternative measures only if they satisfy the Commission that these better mitigate risk (s.13(2) ONSA).\nApplies to: platform\, application s-service-provider\, content-applications-service-provider\nWho: licensed service providers (MCMC has applied the ONSA regime to platforms with mo re than 8 million Malaysian users)\nPenalty: Enforcement under the Online Safety Act 2025 including financial penalties reported up to RM10 millio n for non-compliance with duties under the Act\nSource says: "Date of Pub lication: 22 May 2026 / Date of Enforcement: 1 June 2026 (cover page of t he Code\, MCMC)."\nSource: https://www.mcmc.gov.my/skmmgovmy/media/Genera l/pdf2/ONSA-Risk-Mitigation-Code_1.pdf\nVerified: Aug 30\, 2026\nhttps:// agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Malaysia: Licensed service providers must label synthetic medi a: generated or manipulated images\, audio or video that closely resemble real persons\, objects\, places\, entities or events and are likely to f alsely appear authentic must be clearly distinguishable through prominent labels or markings on the service's online interface (para. 4.2.4(d)\, ' Safe Design of the Service'). The same Code also requires providers to te st and adapt algorithmic and recommender systems against exposure to harm ful content (para. 4.2.4(c)) and to run a documented harmful-content risk assessment reviewed at least annually. END:VALARM END:VEVENT BEGIN:VEVENT UID:uk|automated-vehicles-act-2024-part-5-permits-for-automated-passenger- ser|the-permit-regime-for-automated-passenger-services-goes-live|2026-05- 15@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260515 DTEND;VALUE=DATE:20260516 SUMMARY:United Kingdom: The permit regime for automated passenger services goes live: self-driving taxi\, private-hire and bus-like services must o perate under a permit granted by the appropriate national authority DESCRIPTION:The permit regime for automated passenger services goes live: self-driving taxi\, private-hire and bus-like services must operate under a permit granted by the appropriate national authority\nAutomated Vehicl es Act 2024\, Part 5 (permits for automated passenger services)\, commenc ed by SI 2026/437 (Commencement No. 2) Regulations 2026\, together with t he Automated Vehicles (Permits for Automated Passenger Services) Regulati ons 2026 (SI 2026/439)\nSI 2026/437 commenced the whole of Part 5 of the Automated Vehicles Act 2024 on 15 May 2026\, except s.84 (civil sanctions for infringements of the permit scheme) and s.89(8)(b) and (10). Section 82 empowers the appropriate national authority to grant permits for 'aut omated passenger services' - the carrying of passengers in a road vehicle 'designed or adapted to travel autonomously' or used in a trial to devel op such vehicles - and the permit must specify areas\, vehicles\, validit y period and conditions. The procedural regime (applications\, notices\, review) is in SI 2026/439\, which came into force the same day. Section 8 3 disapplies taxi\, private hire vehicle and bus legislation for services covered by a permit.\nApplies to: automated passenger service operators\ , self-driving taxi and PHV operators\, bus-like autonomous shuttle opera tors\, no-user-in-charge operators\nWho: all\nPenalty: Permit conditions are binding on the permit holder\; note that s.84 (civil sanctions for in fringements of the permit scheme) is expressly NOT yet commenced\nSource says: "The following provisions of the Automated Vehicles Act 2024\, so f ar as not already in force\, come into force on 15th May 2026— (a) Part 5 (permits for automated passenger services)\, except— (i) section 84 (civil sanctions for infringements of the permit scheme)\, and (ii) secti on 89(8)(b) and (10) (procedural and administrative matters)\; (b) sectio n 93 (provision of information about traffic regulation measures)."\nSour ce: https://www.legislation.gov.uk/uksi/2026/437/regulation/2/made\nVerif ied: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Registration TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United Kingdom: The permit regime for automated passenger serv ices goes live: self-driving taxi\, private-hire and bus-like services mu st operate under a permit granted by the appropriate national authority END:VALARM END:VEVENT BEGIN:VEVENT UID:uk|the-data-protection-act-2018-code-of-practice-on-artificial-intelli gen|statutory-duty-on-the-information-commissioner-to-prepare-a|2026-05-1 2@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260512 DTEND;VALUE=DATE:20260513 SUMMARY:United Kingdom: Statutory duty on the Information Commissioner to prepare a code of practice on processing personal data for developing and using AI and for automated decision-making takes effect DESCRIPTION:Statutory duty on the Information Commissioner to prepare a co de of practice on processing personal data for developing and using AI an d for automated decision-making takes effect\nThe Data Protection Act 201 8 (Code of Practice on Artificial Intelligence and Automated Decision-Mak ing) Regulations 2026\, SI 2026/425\nRegulation 2 provides: 'The Commissi oner must prepare an appropriate code of practice giving guidance as to g ood practice in the processing of personal data under the relevant data p rotection legislation in relation to (a) developing and using artificial intelligence\, and (b) automated decision-making.' This is the secondary legislation the Government committed to during passage of the Data (Use a nd Access) Act 2025. Once issued under the Data Protection Act 2018 proce dure\, the code is a statutory code: the Commissioner must take it into a ccount when exercising regulatory functions and it is admissible in evide nce in legal proceedings\, so it will in practice bind controllers develo ping or deploying AI.\nApplies to: Information Commissioner (duty holder) \, data controllers developing or using AI\, deployers of automated decis ion-making\nWho: all\nPenalty: None directly\; the resulting statutory co de will be taken into account by the ICO in enforcement and is admissible in evidence in court and tribunal proceedings\nSource says: "Made 16th A pril 2026 / Laid before Parliament 21st April 2026 / Coming into force 12 th May 2026"\nSource: https://www.legislation.gov.uk/uksi/2026/425/made\n Verified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United Kingdom: Statutory duty on the Information Commissioner to prepare a code of practice on processing personal data for developing and using AI and for automated decision-making takes effect END:VALARM END:VEVENT BEGIN:VEVENT UID:vn|decree-no-142-2026-nd-cp-of-30-april-2026-detailing-articles-and-im ple|operative-implementing-rules-for-the-ai-law-take-effect-risk|2026-05- 01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260501 DTEND;VALUE=DATE:20260502 SUMMARY:Vietnam: Operative implementing rules for the AI Law take effect: risk classification of AI systems into high/medium/low\, conformity asses sment for high-risk AI systems\, accountability/transparency and incident -handling duties\, registration on the one-stop AI portal and the nationa l AI systems database\, and the regulatory sandbox DESCRIPTION:Operative implementing rules for the AI Law take effect: risk classification of AI systems into high/medium/low\, conformity assessment for high-risk AI systems\, accountability/transparency and incident-hand ling duties\, registration on the one-stop AI portal and the national AI systems database\, and the regulatory sandbox\nDecree No. 142/2026/ND-CP of 30 April 2026 detailing articles and implementation measures of the La w on Artificial Intelligence (Nghị định 142/2026/NĐ-CP)\n8 chapters \, 46 articles plus forms. Applies to AI system providers\, developers\, deployers and users\, and to Vietnamese and foreign bodies\, organisation s and individuals engaged in AI activity in Vietnam. Providers of medium- and high-risk AI systems must notify their risk classification before pu tting the system into use. Serious-incident reporting: preliminary report within 72 hours (emergency) or 5 working days (other serious incidents)\ , final report within 15 days of the preliminary report.\nApplies to: dev eloper\, provider\, deployer\, user\nWho: all\nPenalty: Administrative me asures under the AI Law\; suspension/termination powers for systems posin g serious risk\nSource says: "Ngày ban hành: 30-04-2026\; Ngày có hi ệu lực: 01-05-2026\; Người ký: Hồ Quốc Dũng. // 'Date issued : 30 April 2026\; Effective date: 1 May 2026\; Signed by: Hồ Quốc Dũ ng.' Signed text: https://datafiles.chinhphu.vn/cpp/files/vbpq/2026/4/142 -2026-ndcp.signed.pdf"\nSource: https://vanban.chinhphu.vn/?pageid=27160& docid=218029\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-c alendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Vietnam: Operative implementing rules for the AI Law take effe ct: risk classification of AI systems into high/medium/low\, conformity a ssessment for high-risk AI systems\, accountability/transparency and inci dent-handling duties\, registration on the one-stop AI portal and the nat ional AI systems database\, and the regulatory sandbox END:VALARM END:VEVENT BEGIN:VEVENT UID:vn|decree-no-142-2026-nd-cp-article-18-notification-of-ai-interaction- and|synthetic-content-marking-and-labelling-duty-providers-must|2026-05-0 1@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260501 DTEND;VALUE=DATE:20260502 SUMMARY:Vietnam: Synthetic-content marking and labelling duty: PROVIDERS m ust apply technical solutions so that audio\, image and video outputs are marked in a machine-readable format\; DEPLOYERS must give clear notice a nd an easily recognisable label when releasing to the public content gene rated or edited by AI that could mislead as to the authenticity of an eve nt\, a person or the origin of the content DESCRIPTION:Synthetic-content marking and labelling duty: PROVIDERS must a pply technical solutions so that audio\, image and video outputs are mark ed in a machine-readable format\; DEPLOYERS must give clear notice and an easily recognisable label when releasing to the public content generated or edited by AI that could mislead as to the authenticity of an event\, a person or the origin of the content\nDecree No. 142/2026/ND-CP\, Articl e 18 — notification of AI interaction and labelling of AI-generated con tent (read with the Law on Artificial Intelligence No. 134/2025/QH15)\nSp lit duty between nhà cung cấp (provider) and bên triển khai (deploy er). Machine-readable marking at the provider layer plus human-visible la belling at the distribution layer — the same two-layer design as the EU AI Act Art. 50. The Ministry of Science and Technology publishes and upd ates reference technical guidance on the form of the notification and of the displayed label.\nApplies to: provider\, deployer\nWho: all\nPenalty: Administrative liability under the Law on AI and its implementing decree \; suspension/termination powers for systems posing serious risk\nSource says: "Nghị định có hiệu lực từ ngày 01/5/2026. … nhà cun g cấp phải áp dụng giải pháp kỹ thuật để nội dung đ ầu ra là âm thanh\, hình ảnh\, video được đánh dấu ở đ ịnh dạng máy đọc\; bên triển khai phải thông báo rõ ràng \, gắn nhãn dễ nhận biết khi cung cấp ra công chúng nội du ng do AI tạo ra hoặc chỉnh sửa … có khả năng gây nhầm l ẫn về tính xác thực của sự kiện\, nhân vật hoặc nguồ n gốc nội dung. // 'The Decree takes effect from 1 May 2026. … the provider must apply technical solutions so that output audio\, image and video content is marked in a machine-readable format\; the deployer must give clear notice and an easily recognisable label when providing to the public content generated or edited by AI … that is liable to cause conf usion as to the authenticity of an event\, a person or the origin of the content.'"\nSource: https://khpl.moj.gov.vn/portal/tin-tuc/chi-tiet/gioi- thieu-nghi-inh-so-1422026n-cp-quy-inh-chi-tiet-mot-so-ieu-va-bien-phap-th i-hanh-luat-tri-tue-nhan-tao-th2pbneb5f.html\nVerified: Aug 30\, 2026\nht tps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Vietnam: Synthetic-content marking and labelling duty: PROVIDE RS must apply technical solutions so that audio\, image and video outputs are marked in a machine-readable format\; DEPLOYERS must give clear noti ce and an easily recognisable label when releasing to the public content generated or edited by AI that could mislead as to the authenticity of an event\, a person or the origin of the content END:VALARM END:VEVENT BEGIN:VEVENT UID:us|executive-order-ensuring-a-national-policy-framework-for-artificial -in|commerce-identifies-state-ai-laws-for-challenge-by-the-doj-a|2026-03- 11@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260311 DTEND;VALUE=DATE:20260312 SUMMARY:United States — federal: Commerce identifies state AI laws for c hallenge by the DOJ AI Litigation Task Force\; BEAD funding eligibility t ied to state AI rules DESCRIPTION:Commerce identifies state AI laws for challenge by the DOJ AI Litigation Task Force\; BEAD funding eligibility tied to state AI rules\n Executive Order\, Ensuring a National Policy Framework for Artificial Int elligence (11 Dec 2025)\nNot a business compliance duty but the reason se veral state deadlines below moved. The Task Force challenges state AI law s deemed inconsistent with federal policy\; states with targeted rules ri sk losing BEAD broadband funds.\nApplies to: state government\nSource say s: "Sec. 4. Evaluation of State AI Laws. Within 90 days of the date of th is order\, the Secretary of Commerce\, consistent with the Secretary's au thorities under 47 U.S.C. 902(b)\, shall\, in consultation with the Speci al Advisor for AI and Crypto\, the Assistant to the President for Economi c Policy\, the Assistant to the President for Science and Technology\, an d the Assistant to the President and Counsel to the President\, publish a n evaluation of existing State AI laws that identifies onerous laws that conflict with the policy set forth in section 2 of this order\, as well a s laws that should be referred to the Task Force established pursuant to section 3 of this order."\nSource: https://www.whitehouse.gov/presidentia l-actions/2025/12/eliminating-state-law-obstruction-of-national-artificia l-intelligence-policy/\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-co mpliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Regulator milestone TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: Commerce identifies state AI laws f or challenge by the DOJ AI Litigation Task Force\; BEAD funding eligibili ty tied to state AI rules END:VALARM END:VEVENT BEGIN:VEVENT UID:vn|law-on-artificial-intelligence-no-134-2025-qh15-lu-t-tr-tu-nh-n-t-o |vietnam-s-dedicated-ai-law-enters-into-force-risk-tiered-dut|2026-03-01@ agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260301 DTEND;VALUE=DATE:20260302 SUMMARY:Vietnam: Vietnam's dedicated AI Law enters into force: risk-tiered duties on developers\, providers and deployers of AI systems (risk class ification\, data/testing/monitoring and human-oversight requirements for high-risk uses in finance\, health\, justice\, labour and education\; tra nsparency and marking of AI-generated content) DESCRIPTION:Vietnam's dedicated AI Law enters into force: risk-tiered duti es on developers\, providers and deployers of AI systems (risk classifica tion\, data/testing/monitoring and human-oversight requirements for high- risk uses in finance\, health\, justice\, labour and education\; transpar ency and marking of AI-generated content)\nLaw on Artificial Intelligence No. 134/2025/QH15 (Luật Trí tuệ nhân tạo)\nPassed by the 15th Na tional Assembly at its 10th session on 10 December 2025 by 429/434 deputi es (90.70%). 8 chapters\, 35 articles. First comprehensive Vietnamese sta tute governing research\, development\, provision\, deployment and use of AI systems. Technical detail is delegated to implementing decrees (see D ecree 142/2026/ND-CP).\nApplies to: developer\, provider\, deployer\, use r\nWho: all\nPenalty: Administrative and criminal liability under general Vietnamese law\; detail delegated to implementing decrees\nSource says: "Luật Trí tuệ nhân tạo chính thức có hiệu lực từ hôm n ay\, ngày 01/3/2026. // 'The Law on Artificial Intelligence officially t akes effect from today\, 1 March 2026.' Also: 'Ngày 10/12/2025\, Quốc hội biểu quyết thông qua Luật Trí tuệ nhân tạo' ('On 10 De cember 2025 the National Assembly voted to pass the Law on Artificial Int elligence')."\nSource: https://mst.gov.vn/ai-la-cong-cu-ho-tro-quyet-dinh -cuoi-cung-van-la-con-nguoi-19726030111172663.htm\nVerified: Aug 30\, 202 6\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Vietnam: Vietnam's dedicated AI Law enters into force: risk-ti ered duties on developers\, providers and deployers of AI systems (risk c lassification\, data/testing/monitoring and human-oversight requirements for high-risk uses in finance\, health\, justice\, labour and education\; transparency and marking of AI-generated content) END:VALARM END:VEVENT BEGIN:VEVENT UID:in|information-technology-intermediary-guidelines-and-digital-media-et hic|intermediaries-must-label-and-trace-synthetically-generated|2026-02-2 0@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260220 DTEND;VALUE=DATE:20260221 SUMMARY:India: Intermediaries must label and trace synthetically generated content DESCRIPTION:Intermediaries must label and trace synthetically generated co ntent\nInformation Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules\, 2026\nGazette notification G.S.R. 120(E) d ated 10 February 2026\, in force 20 February 2026. The synthetically gene rated information definition covers audio and visual content — text is excluded — and the draft's 10% surface-area watermark rule did not surv ive into the notified version.\nApplies to: intermediary\, platform\nSour ce says: "Gazette of India\, Extraordinary\, Part II Section 3(i)\, notif ication G.S.R. 120(E) dated 10 February 2026 (gazette doc id CG-DL-E-1002 2026-269993)\; the Information Technology (Intermediary Guidelines and Di gital Media Ethics Code) Amendment Rules\, 2026 come into force on 20 Feb ruary 2026. Rule 2(1)(wa) definition\, verbatim: 'audio\, visual or audio -visual information which is artificially or algorithmically created\, ge nerated\, modified or altered using a computer resource\, in a manner tha t such information appears to be real\, authentic or true and depicts or portrays any individual or event in a manner that is\, or is likely to be perceived as indistinguishable from a natural person or real-world event .' Labelling standard as notified: the label must be 'clearly and promine ntly' displayed - a visible label on visual content and an audio disclosu re on audio content\, 'prominent and clearly noticeable'. Provenance duty : intermediaries must 'embed permanent metadata or unique identifiers to trace the computer resource used to generate or alter the content'. SSMI duty: obtain a user declaration whether content is SGI and 'deploy approp riate technical measures\, including automated tools\, to verify whether the declaration is accurate'."\nSource: https://www.meity.gov.in/static/u ploads/2026/02/550681ab908f8afb135b0ad42816a1c9.pdf\nVerified: Aug 30\, 2 026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:India: Intermediaries must label and trace synthetically gener ated content END:VALARM END:VEVENT BEGIN:VEVENT UID:uk|data-use-and-access-act-2025-s-138-creating-or-requesting-the-creat ion|new-criminal-offences-of-creating-or-requesting-the-creation|2026-02- 06@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260206 DTEND;VALUE=DATE:20260207 SUMMARY:United Kingdom: New criminal offences of creating\, or requesting the creation of\, a purported (deepfake) intimate image of an adult witho ut consent come into force DESCRIPTION:New criminal offences of creating\, or requesting the creation of\, a purported (deepfake) intimate image of an adult without consent c ome into force\nData (Use and Access) Act 2025\, s.138 (creating\, or req uesting the creation of\, purported intimate image of adult)\, commenced by SI 2026/31 (Commencement No. 5) Regulations 2026\nSection 138 DUAA ins erts new sections 66E-66H into the Sexual Offences Act 2003. A 'purported intimate image' is an image made or altered so that it appears to be a p hotograph or film of another person in an intimate state - i.e. AI-genera ted or digitally altered 'nudification' and sexual deepfakes. It is an of fence to create such an image of an adult without consent and without rea sonable belief in consent\, and separately to request its creation. Defen ces of reasonable excuse apply\; courts get deprivation-order powers over the images and equipment.\nApplies to: individuals\, operators of nudifi cation and image-generation tools\, platforms hosting such tools\nWho: al l\nPenalty: Criminal offence: on summary conviction\, imprisonment up to the maximum term for summary offences and/or a fine\; deprivation orders in respect of the image and equipment used\nSource says: "Section 138 of the Data (Use and Access) Act 2025 (creating\, or requesting the creation of\, purported intimate image of adult) comes into force on 6th February 2026."\nSource: https://www.legislation.gov.uk/uksi/2026/31/made\nVerifi ed: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Enforcement TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United Kingdom: New criminal offences of creating\, or request ing the creation of\, a purported (deepfake) intimate image of an adult w ithout consent come into force END:VALARM END:VEVENT BEGIN:VEVENT UID:uk|data-use-and-access-act-2025-s-80-and-schedule-6-automated-decision -ma|new-uk-gdpr-articles-22a-22d-automated-decision-making-regim|2026-02- 05@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260205 DTEND;VALUE=DATE:20260206 SUMMARY:United Kingdom: New UK GDPR Articles 22A-22D automated decision-ma king regime takes effect: controllers taking significant solely-automated decisions must have safeguards in place DESCRIPTION:New UK GDPR Articles 22A-22D automated decision-making regime takes effect: controllers taking significant solely-automated decisions m ust have safeguards in place\nData (Use and Access) Act 2025\, s.80 and S chedule 6 (automated decision-making)\, commenced by SI 2026/82 (Commence ment No. 6 and Transitional and Saving Provisions) Regulations 2026\nSect ion 80 DUAA replaces UK GDPR Article 22 with Articles 22A-22D. A decision is 'based solely on automated processing' if there is no meaningful huma n involvement\; where a significant decision is taken about a data subjec t based solely on automated processing\, the controller must ensure safeg uards consisting of or including measures to provide information about th e decision\, enable representations\, enable human intervention and enabl e the decision to be contested. Commencement Regulation 5 of SI 2026/82 s aves the old Article 22(3) regime for decisions taken before 5 February 2 026\, so the new duties bite on decisions taken from that date.\nApplies to: data controllers\, data processors\, deployers of automated decision systems\nWho: all\nPenalty: UK GDPR / DPA 2018 enforcement: penalty notic es up to the higher maximum (GBP 17.5m or 4% of total annual worldwide tu rnover)\, plus enforcement notices and data subject compensation claims\n Source says: "The following provisions of the 2025 Act\, so far as not al ready in force\, come into force on 5th February 2026— ... (j) section 80 (automated decision-making)\; ... (z8) Schedule 6 (automated decision- making: minor and consequential amendments)."\nSource: https://www.legisl ation.gov.uk/uksi/2026/82/regulation/2/made\nVerified: Aug 30\, 2026\nhtt ps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United Kingdom: New UK GDPR Articles 22A-22D automated decisio n-making regime takes effect: controllers taking significant solely-autom ated decisions must have safeguards in place END:VALARM END:VEVENT BEGIN:VEVENT UID:tw|artificial-intelligence-basic-act-promulgated-by-presidential-order -11|taiwan-s-ai-basic-act-enters-into-force-on-promulgation-20-a|2026-01- 14@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260114 DTEND;VALUE=DATE:20260115 SUMMARY:Taiwan: Taiwan's AI Basic Act enters into force on promulgation: 2 0 articles establishing the 7 governing principles\, a duty that high-ris k AI applications be labelled (Art. 5)\, MODA-led AI risk-classification framework and sectoral management rules with power to restrict or prohibi t unlawful applications (Art. 16)\, and allocation of liability plus a re dress/compensation mechanism for high-risk AI applications (Art. 17) DESCRIPTION:Taiwan's AI Basic Act enters into force on promulgation: 20 ar ticles establishing the 7 governing principles\, a duty that high-risk AI applications be labelled (Art. 5)\, MODA-led AI risk-classification fram ework and sectoral management rules with power to restrict or prohibit un lawful applications (Art. 16)\, and allocation of liability plus a redres s/compensation mechanism for high-risk AI applications (Art. 17)\nArtific ial Intelligence Basic Act (人工智慧基本法)\, promulgated by Presi dential Order 華總一義字第11500001671號 of 14 January 2026\nPassed by the Legislative Yuan at third reading on 23 December 2025 (民國114 年12月23日)\; promulgated and brought into force 14 January 2026. Cent ral competent authority: National Science and Technology Council (國科 會)\; local: municipal and county governments. Art. 18 requires governme nt to complete adaptation of laws and administrative measures within two years of commencement (a PERIOD\, not a stated date — not recorded as a separate row). Art. 6 requires the Executive Yuan to establish a Nationa l AI Strategy Special Committee and adopt a National AI Development Progr amme.\nApplies to: developer\, provider\, deployer\, government\nWho: all \nPenalty: No penal provisions in the Act itself. Enforcement runs throug h sectoral competent authorities: under Art. 16\, where an application br eaches Art. 5 it 'shall be restricted or prohibited in accordance with la ws and regulations' (有第5條違法情事者，應依法令限制或禁 止).\nSource says: "公布日期：民國 115 年 01 月 14 日 … 第 二十條「本法自公布之日起施行。」 // 'Promulgation date: 1 4 January 2026 (ROC year 115) … Article 20: This Act shall come into fo rce from the date of promulgation.'"\nSource: https://law.moj.gov.tw/LawC lass/LawAll.aspx?pcode=H0160093\nVerified: Aug 30\, 2026\nhttps://agenccy .ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Taiwan: Taiwan's AI Basic Act enters into force on promulgatio n: 20 articles establishing the 7 governing principles\, a duty that high -risk AI applications be labelled (Art. 5)\, MODA-led AI risk-classificat ion framework and sectoral management rules with power to restrict or pro hibit unlawful applications (Art. 16)\, and allocation of liability plus a redress/compensation mechanism for high-risk AI applications (Art. 17) END:VALARM END:VEVENT BEGIN:VEVENT UID:qa-qfc|qicdrc-practice-direction-no-1-of-2026-with-accompanying-practi ce-guid|litigants-and-legal-representatives-before-the-qfc-court-and|2026 -01-06@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260106 DTEND;VALUE=DATE:20260107 SUMMARY:Qatar — Qatar Financial Centre (QICDRC courts): Litigants and le gal representatives before the QFC Court and the Regulatory Tribunal must verify the accuracy of AI-assisted submissions\, identify AI-assisted po rtions when required\, and must not input confidential\, privileged or le gally protected information into publicly available AI tools DESCRIPTION:Litigants and legal representatives before the QFC Court and t he Regulatory Tribunal must verify the accuracy of AI-assisted submission s\, identify AI-assisted portions when required\, and must not input conf idential\, privileged or legally protected information into publicly avai lable AI tools\nQICDRC Practice Direction No. 1 of 2026\, with accompanyi ng Practice Guidance on the use of artificial intelligence in court proce edings\nIssued by the Qatar International Court and Dispute Resolution Ce ntre following the judgment in Jonathan David Sheppard v Jillion LLC [202 5] QIC (E) 3. Sets a framework for the responsible\, transparent and disc iplined use of AI tools in proceedings. Legal representatives remain resp onsible for the accuracy of all material submitted regardless of whether AI was used\; where the Court requires it\, they must identify AI-assiste d portions and explain the steps taken to verify accuracy and ensure comp liance with professional and ethical standards. Advance disclosure of AI use is not automatically mandatory unless the Court requires it.\nApplies to: litigants\, legal representatives\, counsel\nWho: all\nPenalty: Cour t sanctions in proceedings — exclusion of AI-generated evidence\, adver se costs and professional/ethical consequences for legal representatives\ ; no separate monetary penalty schedule\nSource says: "The Qatar Internat ional Court and Dispute Resolution Centre (QICDRC) has issued Practice Di rection No. 1 of 2026\, together with accompanying Practice Guidance on t he use of artificial intelligence in court proceedings\, setting out a co mprehensive framework governing the responsible\, transparent\, and disci plined use of AI tools before the QFC Court and the Regulatory Tribunal." \nSource: https://www.qicdrc.gov.qa/media-center/news/qicdrc-issues-pract ice-direction-no-1-2026-and-practice-guidance-use-artificial\nVerified: A ug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Qatar — Qatar Financial Centre (QICDRC courts): Litigants an d legal representatives before the QFC Court and the Regulatory Tribunal must verify the accuracy of AI-assisted submissions\, identify AI-assiste d portions when required\, and must not input confidential\, privileged o r legally protected information into publicly available AI tools END:VALARM END:VEVENT BEGIN:VEVENT UID:vn|law-on-digital-technology-industry-no-71-2025-qh15-lu-t-c-ng-nghi-p -c|first-binding-vietnamese-ai-transparency-duties-took-effect|2026-01-01 @agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260101 DTEND;VALUE=DATE:20260102 SUMMARY:Vietnam: First binding Vietnamese AI transparency duties took effe ct: an AI system interacting directly with a person must tell the user th ey are interacting with an AI system\, and digital products on the Minist er's list of AI-generated digital products must carry an identifying mark readable by a person or a machine DESCRIPTION:First binding Vietnamese AI transparency duties took effect: a n AI system interacting directly with a person must tell the user they ar e interacting with an AI system\, and digital products on the Minister's list of AI-generated digital products must carry an identifying mark read able by a person or a machine\nLaw on Digital Technology Industry No. 71/ 2025/QH15 (Luật Công nghiệp công nghệ số)\, Chapter IV — Arti ficial Intelligence\, Art. 44\nLaw passed 14 June 2025\, published in Cô ng báo issues 965+966 of 24 July 2025\, signed by National Assembly Chai rman Trần Thanh Mẫn. 6 chapters\, 51 articles\; Chapter IV (Arts. 41- 45) covered AI. The Minister of Science and Technology was to issue the L ist of AI-generated digital products and to inspect compliance. SUPERSEDE D: Art. 34 of the Law on Artificial Intelligence No. 134/2025/QH15 repeal s Chapter IV of Law 71/2025/QH15 (together with Art. 3(9)\, Art. 4(7)\, A rt. 12(6) and Art. 34(2)(dd)) with effect from 1 March 2026\; equivalent transparency and AI-content labelling duties continue under the AI Law.\n Applies to: provider\, deployer\nWho: all\nPenalty: Administrative inspec tion by the Ministry of Science and Technology\; general administrative l iability\nSource says: "Ngày ban hành: 14/06/2025\; Ngày hiệu lực: 01/01/2026 (Công báo số 965 + 966\, ngày 24/7/2025\; người ký: Trần Thanh Mẫn). // 'Issued 14 June 2025\; effective 1 January 2026.' "\nSource: https://congbao.chinhphu.vn/van-ban/luat-so-71-2025-qh15-45555 .htm\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Vietnam: First binding Vietnamese AI transparency duties took effect: an AI system interacting directly with a person must tell the use r they are interacting with an AI system\, and digital products on the Mi nister's list of AI-generated digital products must carry an identifying mark readable by a person or a machine END:VALARM END:VEVENT BEGIN:VEVENT UID:ca-on|employment-standards-act-2000-s-o-2000-c-41-part-iii-1-job-posti ngs-s|employers-must-disclose-in-a-publicly-advertised-job-posting|2026-0 1-01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260101 DTEND;VALUE=DATE:20260102 SUMMARY:Canada — Ontario: Employers must disclose in a publicly advertis ed job posting that artificial intelligence is used to screen\, assess or select applicants DESCRIPTION:Employers must disclose in a publicly advertised job posting t hat artificial intelligence is used to screen\, assess or select applican ts\nEmployment Standards Act\, 2000 (S.O. 2000\, c. 41)\, Part III.1 'Job Postings'\, s. 8.4 — added by the Working for Workers Four Act\, 2024 (S.O. 2024\, c. 3)\, Schedule 2\, s. 2(1)\; supported by O. Reg. 476/24 ( Rules and Exemptions re Job Postings)\ns. 8.4(1) is a positive disclosure duty: if AI is used anywhere in screening\, assessment or selection for the advertised position\, a statement disclosing that use must appear in the posting itself. O. Reg. 476/24 defines 'artificial intelligence' in O ECD terms and exempts employers with fewer than 25 employees on the day t he posting is posted\, general recruitment campaigns and help-wanted sign s that do not advertise a specific position\, internal-only postings\, an d postings for work performed outside Ontario. This is the first in-force statutory AI-hiring disclosure duty in Canada.\nApplies to: employer\, d eployer\nWho: employers with 25 or more employees on the day the posting is posted (O. Reg. 476/24 s. 1 exempts employers with fewer than 25 emplo yees)\nPenalty: Enforced under the Employment Standards Act\, 2000: emplo yment standards officers may issue compliance orders and notices of contr avention carrying prescribed administrative penalties\, and contravention of the Act is an offence prosecutable under Part XXV of the ESA.\nSource says: "Use of artificial intelligence — 8.4 (1) Every employer who adv ertises a publicly advertised job posting and who uses artificial intelli gence to screen\, assess or select applicants for the position shall incl ude in the posting a statement disclosing the use of the artificial intel ligence. 2024\, c. 3\, Sched. 2\, s. 2 (1). [Section Amendments with date in force (d/m/y): 2024\, c. 3\, Sched. 2\, s. 2 (1) - 01/01/2026]"\nSour ce: https://www.ontario.ca/laws/statute/00e41\nVerified: Aug 30\, 2026\nh ttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Canada — Ontario: Employers must disclose in a publicly adve rtised job posting that artificial intelligence is used to screen\, asses s or select applicants END:VALARM END:VEVENT BEGIN:VEVENT UID:ca-qc|indication-d-application-ia-ri-2025-003-op-mesures-applicables-l ors-de|binding-generative-ai-measures-apply-to-quebec-public-bodies|2025- 12-05@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20251205 DTEND;VALUE=DATE:20251206 SUMMARY:Canada — Quebec: Binding generative-AI measures apply to Quebec public bodies: organizational AI governance structure\, risk and complian ce review before deployment\, data-protection and confidentiality rules\, mandatory preference for secure private GenAI systems over public ones\, and lifecycle quality/security controls DESCRIPTION:Binding generative-AI measures apply to Quebec public bodies: organizational AI governance structure\, risk and compliance review befor e deployment\, data-protection and confidentiality rules\, mandatory pref erence for secure private GenAI systems over public ones\, and lifecycle quality/security controls\nIndication d'application IA-RI-2025-003-OP — « Mesures applicables lors de l'utilisation de l'intelligence artificie lle générative » (formulated by the Dirigeant principal de l'informati on under LGGRI\, chapitre G-1.03\, art. 7)\; complements arrêté minist ériel 2025-02 of 3 December 2025 (Énoncé de principes pour une utilisa tion responsable de l'IA par les organismes publics)\nAn 'indication d'ap plication' is a written instruction with which public bodies MUST comply under the Act respecting the governance and management of the information resources of public bodies and government enterprises (chapter G-1.03). Status field on the instrument reads 'En vigueur'. It applies to the publ ic bodies listed in s.2 of the LGGRI — ministries\, government agencies \, health and social services bodies\, school service centres and Crown c orporations. It also repeals the earlier IA-RI-2025-001-OP suspension of generative-AI virtual assistants\, and sits alongside standing prohibitio ns (DeepSeek assistants\, Microsoft Teams voice/face enrolment).\nApplies to: deployer\, user\nWho: all bodies covered by art. 2 LGGRI\nPenalty: N o monetary penalty in the instrument. Compliance is mandatory under the L GGRI (c. G-1.03)\; the Dirigeant principal de l'information may under art . 24 exempt a body in whole or in part and set conditions. Enforcement is administrative through the Ministère de la Cybersécurité et du Numér ique.\nSource says: "Statut : En vigueur — No de référence : IA-RI-20 25-003-OP — Organismes visés : Organismes publics — Référence lég ale : LGGRI (chapitre G-1.03)\, art. 7 — Date de formulation : 2025-12- 05 — Date d'entrée en vigueur : 2025-12-05"\nSource: https://cdn-conte nu.quebec.ca/cdn-contenu/adm/min/cybersecurite_numerique/Publications/Str ategie_IA/Indication_app_IAG.pdf\nVerified: Aug 30\, 2026\nhttps://agencc y.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Law applies TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Canada — Quebec: Binding generative-AI measures apply to Que bec public bodies: organizational AI governance structure\, risk and comp liance review before deployment\, data-protection and confidentiality rul es\, mandatory preference for secure private GenAI systems over public on es\, and lifecycle quality/security controls END:VALARM END:VEVENT BEGIN:VEVENT UID:nz|biometric-processing-privacy-code-2025-issued-by-the-privacy-commis sio|binding-code-of-practice-for-automated-biometric-processing|2025-11-0 3@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20251103 DTEND;VALUE=DATE:20251104 SUMMARY:New Zealand: Binding code of practice for automated biometric proc essing comes into force for NEW biometric processing: 13 modified privacy rules including a proportionality assessment before collection (Rule 1 — lawful purpose\, necessity\, effectiveness and no less privacy-intrus ive alternative reasonably available)\, transparency/notification duties\ , restrictions on biometric categorisation\, retention limits and cross-b order safeguards DESCRIPTION:Binding code of practice for automated biometric processing co mes into force for NEW biometric processing: 13 modified privacy rules in cluding a proportionality assessment before collection (Rule 1 — lawful purpose\, necessity\, effectiveness and no less privacy-intrusive altern ative reasonably available)\, transparency/notification duties\, restrict ions on biometric categorisation\, retention limits and cross-border safe guards\nBiometric Processing Privacy Code 2025\, issued by the Privacy Co mmissioner under the Privacy Act 2020\nApplies to organisations that coll ect biometric information in an AUTOMATED process to verify\, identify or categorise people. 'Biometric categorisation' expressly covers inference systems — e.g. using a biometric system to infer someone's emotions fr om their voice or to estimate their age from their face. Health agencies carrying out biometric processing to deliver health services are outside the Code.\nApplies to: deployer\, provider\nWho: all agencies as defined in the Privacy Act 2020\nPenalty: Breach of a code rule is treated as an interference with privacy under the Privacy Act 2020 — complaints to th e Privacy Commissioner\, compliance notices\, and Human Rights Review Tri bunal proceedings with damages. OPC applies its Compliance and Regulatory Action Framework.\nSource says: "The Biometric Processing Privacy Code 2 025 (BPPC) was issued on 21 July 2025. … The BPPC came into force on 3 November 2025\, but agencies already using biometrics had a nine-month gr ace period to move to the new set of rules."\nSource: https://www.privacy .org.nz/privacy-principles/codes-of-practice/biometric-processing-privacy -code/\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calenda r/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:New Zealand: Binding code of practice for automated biometric processing comes into force for NEW biometric processing: 13 modified pri vacy rules including a proportionality assessment before collection (Rule 1 — lawful purpose\, necessity\, effectiveness and no less privacy-int rusive alternative reasonably available)\, transparency/notification duti es\, restrictions on biometric categorisation\, retention limits and cros s-border safeguards END:VALARM END:VEVENT BEGIN:VEVENT UID:ng|nigeria-data-protection-act-ndp-act-2023-general-application-and-im ple|data-controllers-and-data-processors-deploying-artificial-in|2025-09- 19@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20250919 DTEND;VALUE=DATE:20250920 SUMMARY:Nigeria: Data controllers and data processors deploying Artificial Intelligence (and other emerging technologies) to process personal data must set and document technical and organisational parameters\, carry out a DPIA\, test in low-risk environments\, assess disparate outcomes\, re- test\, and put continuous monitoring in place — with the parameters and DPIA filed with the Commission as part of NDP Act Compliance Audit Retur ns DESCRIPTION:Data controllers and data processors deploying Artificial Inte lligence (and other emerging technologies) to process personal data must set and document technical and organisational parameters\, carry out a DP IA\, test in low-risk environments\, assess disparate outcomes\, re-test\ , and put continuous monitoring in place — with the parameters and DPIA filed with the Commission as part of NDP Act Compliance Audit Returns\nN igeria Data Protection Act (NDP Act) 2023 — General Application and Imp lementation Directive (GAID) 2025\, ref. NDPC/NDP ACT-GAID/01/2025\, Arti cle 43 (Emerging Technologies)\nArticle 43(1) applies to any controller/p rocessor who 'deploys or intends to deploy Emerging Technologies (ETs) su ch as Artificial Intelligence\, Internet of Things and Blockchain for the purposes of processing personal data'. Art 43(2) requires parameters tha t take account of the right not to be subject to a decision solely based on automated processes or algorithms\, the right to be forgotten\, sensit ive-data and child safeguards\, cross-border flows\, and privacy by desig n and default. Art 43(3) makes the parameters filable with the Commission as part of the Compliance Audit Return. Art 43(4) requires a DPIA (accou nting for disparate outcomes and Data Subjects' Vulnerability Indexes)\, data anonymisation suitability\, low-risk-environment testing\, retooling and re-testing until satisfactory or discarding the tool for unmitigable privacy risk\, and continuous monitoring once deemed safe. Art 44(3) req uires controllers to 'refrain from or cease the use of ET systems that ar e impossible to operate in compliance with international human rights law '. GAID repealed the NDPR 2019 and the NDPR Implementation Framework 2020 .\nApplies to: data controller\, data processor\, data controllers and pr ocessors of major importance\, Data Protection Officer\nWho: all\; heavie r registration and audit duties for Data Controllers/Processors of Major Importance\nPenalty: Enforcement under the NDP Act 2023: sanctions up to NGN 10\,000\,000 or 2% of annual gross revenue in the preceding year for data controllers of major importance (NGN 2\,000\,000 or 2% for others)\, plus remediation orders and compensation\nSource says: "Issued under my hand this 20th Day of March\, 2025"\nSource: https://ndpc.gov.ng/wp-conte nt/uploads/2025/07/NDP-ACT-GAID-2025-MARCH-20TH.pdf\nVerified: Aug 30\, 2 026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Nigeria: Data controllers and data processors deploying Artifi cial Intelligence (and other emerging technologies) to process personal d ata must set and document technical and organisational parameters\, carry out a DPIA\, test in low-risk environments\, assess disparate outcomes\, re-test\, and put continuous monitoring in place — with the parameters and DPIA filed with the Commission as part of NDP Act Compliance Audit R eturns END:VALARM END:VEVENT BEGIN:VEVENT UID:cn|measures-for-labelling-ai-generated-synthetic-content-mandatory-sta nda|explicit-and-implicit-labelling-of-ai-generated-text-images|2025-09-0 1@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20250901 DTEND;VALUE=DATE:20250902 SUMMARY:China: Explicit and implicit labelling of AI-generated text\, imag es\, audio\, video and virtual scenes DESCRIPTION:Explicit and implicit labelling of AI-generated text\, images\ , audio\, video and virtual scenes\nMeasures for Labelling AI-Generated S ynthetic Content + mandatory standard GB 45438-2025\nProviders of generat ion and synthesis services must attach both visible labels and embedded m etadata\; distribution platforms must verify and surface them. In force s ince 1 September 2025 — listed as the standing baseline every later Chi nese AI rule builds on.\nApplies to: provider\, platform\nSource says: "M easures\, Article 14 (verbatim): '本办法自2025年9月1日起施行' = 'These Measures shall come into force on 1 September 2025.' Article 3 de fines the two labelling modes and the covered content: 显式标识 = a l abel 'presented by way of text\, sound\, graphics etc. and able to be cle arly perceived by users'\; 隐式标识 = a label 'added by technical mea ns into the data of the generated or synthetic content file\, not readily perceptible to users'\; covered content is '文本、图片、音频、 视频、虚拟场景等信息' = 'text\, images\, audio\, video\, virtua l scenes and other information'. Article 13 (enforcement): '由网信、 电信、公安和广播电视等有关主管部门依据职责，按照 有关法律、行政法规、部门规章的规定予以处理' = 'the c ompetent cyberspace\, telecommunications\, public security and radio-and- television authorities shall handle it according to their duties and in a ccordance with relevant laws\, administrative regulations and departmenta l rules.' GB 45438-2025\, per the official SAMR national standards databa se: 标准号 GB 45438-2025\, 中文标准名称 '网络安全技术 人 工智能生成合成内容标识方法' (Cybersecurity technology - Labe lling method for AI-generated synthetic content)\, 标准性质 强制性 (MANDATORY)\, 发布日期 2025-02-28\, 实施日期 2025-09-01."\nSourc e: https://www.cac.gov.cn/2025-03/14/c_1743654684782215.htm\nVerified: Au g 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:China: Explicit and implicit labelling of AI-generated text\, images\, audio\, video and virtual scenes END:VALARM END:VEVENT BEGIN:VEVENT UID:ca-ab|protection-of-privacy-act-popa-sa-2024-c-p-27-5-s-5-2-d-and-the- protec|public-bodies-must-state-in-their-collection-notice-when-per|2025- 06-11@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20250611 DTEND;VALUE=DATE:20250612 SUMMARY:Canada — Alberta: Public bodies must state in their collection n otice when personal information being collected will be used in an automa ted system to generate content or to make decisions\, recommendations or predictions DESCRIPTION:Public bodies must state in their collection notice when perso nal information being collected will be used in an automated system to ge nerate content or to make decisions\, recommendations or predictions\nPro tection of Privacy Act (POPA)\, SA 2024\, c. P-27.5\, s. 5(2)(d)\, and th e Protection of Privacy Regulation / Protection of Privacy (Ministerial) Regulation\nPOPA replaced Alberta's Freedom of Information and Protection of Privacy Act for the public sector. s. 5(2)(d) makes AI/automated-syst em use a mandatory element of the collection notice given to individuals. Alberta's own guidance adds that where an automated system is actually u sed to make decisions\, the public body must make every reasonable effort to ensure the accuracy and completeness of the personal information and retain it for at least one year after use. The Act applies from proclamat ion forward: personal information collected before that date may still be used in AI/automated systems\, but collection notices must be updated as new information is collected or programs are revised.\nApplies to: deplo yer\, controller\nWho: all public bodies designated under the Access to I nformation Act s. 1(t) and the Designation of Public Bodies Regulation\nP enalty: POPA carries what the Government of Alberta describes as the stro ngest penalties for privacy violations in Canada\; offences under the Act are punishable by fines\, and the Information and Privacy Commissioner o f Alberta has investigation and order powers.\nSource says: "The Protecti on of Privacy Act (POPA) and regulation came into force June 11\, 2025. U pon proclamation\, the Freedom of Information and Protection of Privacy A ct was repealed."\nSource: https://www.alberta.ca/about-the-protection-of -privacy-act\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-c alendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Canada — Alberta: Public bodies must state in their collecti on notice when personal information being collected will be used in an au tomated system to generate content or to make decisions\, recommendations or predictions END:VALARM END:VEVENT BEGIN:VEVENT UID:sg|elections-integrity-of-online-advertising-amendment-act-2024-act-34 -of|criminal-ban-during-an-election-period-on-publishing-boostin|2025-01- 22@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20250122 DTEND;VALUE=DATE:20250123 SUMMARY:Singapore: Criminal ban\, during an election period\, on publishin g\, boosting\, sharing or reposting online election advertising that cont ains a realistic but false digitally generated or manipulated depiction o f a candidate saying or doing something they did not say or do\; Returnin g Officer may issue corrective directions to individuals\, social media s ervices and internet access service providers to take down or block such content DESCRIPTION:Criminal ban\, during an election period\, on publishing\, boo sting\, sharing or reposting online election advertising that contains a realistic but false digitally generated or manipulated depiction of a can didate saying or doing something they did not say or do\; Returning Offic er may issue corrective directions to individuals\, social media services and internet access service providers to take down or block such content \nElections (Integrity of Online Advertising) (Amendment) Act 2024 (Act 3 4 of 2024)\, inserting s.61MA into the Parliamentary Elections Act 1954 a nd an equivalent provision into the Presidential Elections Act 1991\; com menced by Commencement Notification 2025 (S 47/2025)\nThe prohibition bit es from the issue of the Writ of Election until the close of polling. Ele ments of the offence include that 'the representation was created wholly or partly with content that was generated or manipulated using digital me ans' and that it is 'realistic enough such that it is likely that some me mbers of the general public would ... reasonably believe that the candida te said or did that thing'.\nApplies to: individual\, platform\, social-m edia-service\, internet-access-service-provider\nWho: all\nPenalty: Crimi nal offence\; Returning Officer corrective directions with offences for n on-compliance (Parliamentary Elections Act 1954 as amended)\nSource says: "2. The Elections (Integrity of Online Advertising) (Amendment) Act 2024 comes into operation on 22 January 2025. — Made on 16 January 2025. LE O YIP\, Permanent Secretary\, Prime Minister's Office\, Singapore. [First published in the Government Gazette\, Electronic Edition\, on 20 January 2025 at 5 pm.]"\nSource: https://www.eld.gov.sg/gazettes/2025/Elections% 20(Integrity%20of%20Online%20Advertising)%20(Amendment)%20Act%202024%20(C ommencement)%20Notification%202025.pdf\nVerified: Aug 30\, 2026\nhttps:// agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Enforcement TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Singapore: Criminal ban\, during an election period\, on publi shing\, boosting\, sharing or reposting online election advertising that contains a realistic but false digitally generated or manipulated depicti on of a candidate saying or doing something they did not say or do\; Retu rning Officer may issue corrective directions to individuals\, social med ia services and internet access service providers to take down or block s uch content END:VALARM END:VEVENT BEGIN:VEVENT UID:ca-qc|loi-25-act-to-modernize-legislative-provisions-as-regards-the-pr otecti|automated-decision-making-transparency-notify-the-individual|2023- 09-22@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20230922 DTEND;VALUE=DATE:20230923 SUMMARY:Canada — Quebec: Automated decision-making transparency: notify the individual when a decision is based exclusively on automated processi ng of personal information\, and on request disclose the personal informa tion used\, the reasons and the principal factors and parameters that led to the decision\, plus the right to have that information corrected and to submit observations to a member of staff able to review the decision DESCRIPTION:Automated decision-making transparency: notify the individual when a decision is based exclusively on automated processing of personal information\, and on request disclose the personal information used\, the reasons and the principal factors and parameters that led to the decisio n\, plus the right to have that information corrected and to submit obser vations to a member of staff able to review the decision\nLoi 25 — Act to modernize legislative provisions as regards the protection of personal information (assented 22 September 2021\, LQ 2021\, c. 25)\, inserting s . 12.1 into the Act respecting the protection of personal information in the private sector (RLRQ c. P-39.1) and s. 65.2 into the Act respecting a ccess to documents held by public bodies (RLRQ c. A-2.1)\nBroader than GD PR Art. 22: Quebec's duty attaches to ANY decision based exclusively on a utomated processing of personal information\, with no legal-effects or si gnificant-effects threshold and no exemption list. It binds every person carrying on an enterprise in Quebec (s. 12.1 P-39.1) and\, in parallel\, every public body (s. 65.2 Act respecting access). The Government of Queb ec's own guidance states a decision is 'exclusively automated' where no n atural person exercised meaningful control\, and that minor human interve ntion with no real effect on the outcome does not take the decision outsi de the rule.\nApplies to: deployer\, controller\nWho: no size threshold\; applies to any person carrying on an enterprise\nPenalty: Administrative monetary penalties up to CAD 10\,000\,000 or 2% of worldwide turnover (w hichever is greater) and penal fines up to CAD 25\,000\,000 or 4% of worl dwide turnover\, under the Law 25 enforcement provisions in force since 2 2 September 2023\; plus a private right of action for punitive damages fo r intentional or grossly negligent breaches.\nSource says: "Date de la sa nction : 22 septembre 2021 — Chapitre dans le Recueil annuel des lois d u Québec : 2021\, chapitre 25 — Mode d'entrée en vigueur : le 22 sept embre 2023\, sauf exceptions"\nSource: https://www.assnat.qc.ca/fr/travau x-parlementaires/projets-loi/projet-loi-64-42-1.html\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Canada — Quebec: Automated decision-making transparency: not ify the individual when a decision is based exclusively on automated proc essing of personal information\, and on request disclose the personal inf ormation used\, the reasons and the principal factors and parameters that led to the decision\, plus the right to have that information corrected and to submit observations to a member of staff able to review the decisi on END:VALARM END:VEVENT BEGIN:VEVENT UID:ae-difc|difc-data-protection-regulations-consolidated-version-no-2-reg ulation|deployers-and-operators-of-autonomous-semi-autonomous-system|2023 -09-01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20230901 DTEND;VALUE=DATE:20230902 SUMMARY:United Arab Emirates — DIFC: Deployers and Operators of autonomo us/semi-autonomous systems (incl. AI and machine-learning systems) that p rocess personal data must meet Regulation 10 design\, transparency\, acco untability and high-risk requirements DESCRIPTION:Deployers and Operators of autonomous/semi-autonomous systems (incl. AI and machine-learning systems) that process personal data must m eet Regulation 10 design\, transparency\, accountability and high-risk re quirements\nDIFC Data Protection Regulations\, Consolidated Version No. 2 — Regulation 10 (Personal Data Processed Through Autonomous and Semi-A utonomous Systems)\, made under DIFC Data Protection Law No. 5 of 2020\nR egulation 10 is the first binding AI-specific data protection instrument in the MEASA region. Reg 10.3.2 prohibits commercial use of a System to p rocess personal data unless it processes only for human-defined or human- approved purposes and is designed in compliance with Reg 10.3.1 (ethical\ , fair\, transparent\, secure\, accountable). Reg 10.3.3 adds conditions for High Risk Processing Activities\, including appointment of an Autonom ous Systems Officer (ASO). Reg 10.2.2 requires notice to individuals abou t the underlying technology. General certification requirements are left to future Commissioner guidance and were still outstanding as of January 2026.\nApplies to: Deployer\, Operator\, Provider\, Controller\, Processo r\nWho: all\nPenalty: Fines under the DIFC Data Protection Law No. 5 of 2 020 fines schedule\, as revised by DIFC Laws Amendment Law No. 1 of 2025: USD 50\,000 for failing to comply with the general Article 9 processing requirements (which Regulation 10.2.1 expressly applies to System process ing) and USD 50\,000 for failing to carry out a DPIA prior to High Risk P rocessing Activities (raised from USD 20\,000). The Commissioner may also issue directions and enforcement notices\, and since 15 July 2025 data s ubjects have a direct right of action in the DIFC Courts for compensation including non-financial damage such as distress.\nSource says: "In force on 1 September 2023"\nSource: https://www.dataguidance.com/sites/default /files/data_protection_regualtions_final.pdf\nVerified: Aug 30\, 2026\nht tps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United Arab Emirates — DIFC: Deployers and Operators of auto nomous/semi-autonomous systems (incl. AI and machine-learning systems) th at process personal data must meet Regulation 10 design\, transparency\, accountability and high-risk requirements END:VALARM END:VEVENT BEGIN:VEVENT UID:us|fedreg|2026-13656|2026-08-31@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260831 DTEND;VALUE=DATE:20260901 SUMMARY:United States — federal: CMS CY2027 hospital outpatient (OPPS/AS C) payment rule DESCRIPTION:CMS CY2027 hospital outpatient (OPPS/ASC) payment rule\nHealth and Human Services Department — Proposed Rule\nCMS CY2027 hospital out patient (OPPS/ASC) payment rule\; section X.B proposes an interim Medicar e payment framework for AI/algorithm-driven clinical software (renaming S aaS to 'Software as a Medical Service'\, new status indicator O1\, 36 HCP CS codes moved to new-technology APCs) and asks for comment on it.\nAppli es to: any interested party\nSource says: "comments_close_on: 2026-08-31 (Federal Register API)"\nSource: https://www.federalregister.gov/document s/2026/07/07/2026-13656/medicare-program-hospital-outpatient-prospective- payment-and-ambulatory-surgical-center-payment\nVerified: Aug 30\, 2026\n https://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Comments close TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: CMS CY2027 hospital outpatient (OPP S/ASC) payment rule END:VALARM END:VEVENT BEGIN:VEVENT UID:us|fedreg|2026-15483|2026-08-31@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260831 DTEND;VALUE=DATE:20260901 SUMMARY:United States — federal: NHTSA seeks comment on interim guidance for 49 CFR part 555 'General Exemption' applications that let… DESCRIPTION:NHTSA seeks comment on interim guidance for 49 CFR part 555 'G eneral Exemption' applications that let…\nTransportation Department — Notice\nNHTSA seeks comment on interim guidance for 49 CFR part 555 'Gen eral Exemption' applications that let automated-driving-system vehicles b e commercially deployed without meeting all FMVSS\, and on how the exempt ion process for ADS vehicles could be improved.\nApplies to: any interest ed party\nSource says: "comments_close_on: 2026-08-31 (Federal Register A PI)"\nSource: https://www.federalregister.gov/documents/2026/07/31/2026-1 5483/av-framework-updates-and-request-for-comments-on-interim-guidance\nV erified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Comments close TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: NHTSA seeks comment on interim guid ance for 49 CFR part 555 'General Exemption' applications that let… END:VALARM END:VEVENT BEGIN:VEVENT UID:us|fedreg|2026-13928|2026-09-08@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260908 DTEND;VALUE=DATE:20260909 SUMMARY:United States — federal: BLS seeks comment on adding a new AI mo dule to the American Time Use Survey DESCRIPTION:BLS seeks comment on adding a new AI module to the American Ti me Use Survey\nLabor Department — Notice\nBLS seeks comment on adding a new AI module to the American Time Use Survey - questions on whether peo ple use AI tools\, for which tasks\, and how AI use varies by demographic /occupational group - to be fielded from January 2027 for two years.\nApp lies to: any interested party\nSource says: "comments_close_on: 2026-09-0 8 (Federal Register API)"\nSource: https://www.federalregister.gov/docume nts/2026/07/10/2026-13928/proposed-information-collection-atus-artificial -intelligence-ai-questions\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/a i-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Comments close TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: BLS seeks comment on adding a new A I module to the American Time Use Survey END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|cyber-resilience-act-regulation-eu-2024-2847|manufacturers-must-rep ort-actively-exploited-vulnerabilities|2026-09-11@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260911 DTEND;VALUE=DATE:20260912 SUMMARY:European Union: Manufacturers must report actively exploited vulne rabilities and severe incidents to ENISA and the national CSIRT DESCRIPTION:Manufacturers must report actively exploited vulnerabilities a nd severe incidents to ENISA and the national CSIRT\nCyber Resilience Act (Regulation (EU) 2024/2847)\nEarly-warning notification within 24 hours to the CSIRT designated as coordinator and to ENISA\, via the Article 16 single reporting platform. Covers products with digital elements\; Articl e 12 explicitly extends to products classified as high-risk AI systems un der Article 6 of the AI Act.\nApplies to: manufacturer\nPenalty: Up to EU R 15 000 000 or 2.5% of total worldwide annual turnover\, whichever is hi gher (Art. 64(2))\nSource says: "This Regulation shall apply from 11 Dece mber 2027. However\, Article 14 shall apply from 11 September 2026 and Ch apter IV (Articles 35 to 51) shall apply from 11 June 2026. (Article 71(2 )\, Entry into force and application)"\nSource: https://eur-lex.europa.eu /eli/reg/2024/2847/oj/eng\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai -compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Reporting TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Manufacturers must report actively exploited v ulnerabilities and severe incidents to ENISA and the national CSIRT END:VALARM END:VEVENT BEGIN:VEVENT UID:us|fedreg|2026-14358|2026-09-14@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260914 DTEND;VALUE=DATE:20260915 SUMMARY:United States — federal: CMS/CDC request for information on mode rnizing the 1988 CLIA clinical-laboratory regulations\, with one of its … DESCRIPTION:CMS/CDC request for information on modernizing the 1988 CLIA c linical-laboratory regulations\, with one of its…\nHealth and Human Ser vices Department — Proposed Rule\nCMS/CDC request for information on mo dernizing the 1988 CLIA clinical-laboratory regulations\, with one of its numbered topics asking specifically how labs use AI and software algorit hms to interpret test results and whether AI should be written into the C LIA rules.\nApplies to: any interested party\nSource says: "comments_clos e_on: 2026-09-14 (Federal Register API)"\nSource: https://www.federalregi ster.gov/documents/2026/07/16/2026-14358/request-for-information-clinical -laboratory-improvement-amendments-of-1988-clia-regulations\nVerified: Au g 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Comments close TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: CMS/CDC request for information on modernizing the 1988 CLIA clinical-laboratory regulations\, with one of i ts… END:VALARM END:VEVENT BEGIN:VEVENT UID:us|fedreg|2026-14327|2026-09-14@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260914 DTEND;VALUE=DATE:20260915 SUMMARY:United States — federal: CMS CY2027 Physician Fee Schedule rule which\, alongside routine payment policy\, runs a large RFI on paying… DESCRIPTION:CMS CY2027 Physician Fee Schedule rule which\, alongside routi ne payment policy\, runs a large RFI on paying…\nHealth and Human Servi ces Department — Proposed Rule\nCMS CY2027 Physician Fee Schedule rule which\, alongside routine payment policy\, runs a large RFI on paying for clinical AI in primary care (AI scribes\, decision support\, AI-enabled Annual Wellness Visits\, whether AI companies may deliver AWV services) a nd proposes two AI-focused MIPS improvement activities plus AI/algorithm- driven 'Software as a Medical Service' lab-analysis payment.\nApplies to: any interested party\nSource says: "comments_close_on: 2026-09-14 (Feder al Register API)"\nSource: https://www.federalregister.gov/documents/2026 /07/16/2026-14327/medicare-and-medicaid-programs-cy-2027-payment-policies -under-the-physician-fee-schedule-and-other\nVerified: Aug 30\, 2026\nhtt ps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Comments close TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: CMS CY2027 Physician Fee Schedule r ule which\, alongside routine payment policy\, runs a large RFI on paying … END:VALARM END:VEVENT BEGIN:VEVENT UID:us|fedreg|2026-17761|2026-09-30@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20260930 DTEND;VALUE=DATE:20261001 SUMMARY:United States — federal: FCC seeks comment on overhauling how th e Universal Service Fund is administered by USAC DESCRIPTION:FCC seeks comment on overhauling how the Universal Service Fun d is administered by USAC\nFederal Communications Commission — Proposed Rule\nFCC seeks comment on overhauling how the Universal Service Fund is administered by USAC (processes\, shot clocks\, audits and recoveries\, operating costs\, USAC's board)\, and asks repeatedly whether AI should b e used to review USF applications\, audits and appeals - and what safegua rds\, governance and privacy protections that would require.\nApplies to: any interested party\nSource says: "comments due 2026-09-30\; reply comm ents 2026-10-30 (Federal Register API)"\nSource: https://www.federalregis ter.gov/documents/2026/08/31/2026-17761/maximizing-efficiencies-in-univer sal-service-administration\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/a i-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Comments close TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: FCC seeks comment on overhauling ho w the Universal Service Fund is administered by USAC END:VALARM END:VEVENT BEGIN:VEVENT UID:us-ct|connecticut-ai-responsibility-and-transparency-act-public-act-26 -15-sb|warn-act-layoff-notices-must-state-whether-the-layoffs-are-r|2026- 10-01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20261001 DTEND;VALUE=DATE:20261002 SUMMARY:United States — Connecticut: WARN Act layoff notices must state whether the layoffs relate to the employer's use of AI or another technol ogical change DESCRIPTION:WARN Act layoff notices must state whether the layoffs relate to the employer's use of AI or another technological change\nConnecticut AI Responsibility and Transparency Act (Public Act 26-15\, SB 5)\nFirst o perative date of the CART Act\, signed 2 June 2026. The Act phases in bet ween October 2026 and January 2028 and covers employment decision tools\, consumer chatbots\, frontier developers\, generative-AI provenance and p latforms used by minors.\nApplies to: employer\nWho: employers subject to the federal WARN Act\nSource says: "Closest verified wording (from concu rring analyses of the session law): employers must give written notice to the Connecticut Department of Labor disclosing "whether the layoffs are related to the employer's use of artificial intelligence or another techn ological change"\, triggered where the layoffs "qualify as a mass layoff or plant closing under the federal Worker Adjustment and Retraining Notif ication (WARN) Act"\, effective October 1\, 2026. The row's source_quote ("Beginning October 1\, 2026\, any employer that serves written notice... ") is a commentator's construction\, not statutory drafting -- Connecticu t public acts carry effective dates in a separate effective-date table\, never as an inline 'Beginning [date]' clause."\nSource: https://www.cga.c t.gov/2026/act/pa/pdf/2026PA-00015-R00SB-00005-PA.pdf\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — Connecticut: WARN Act layoff notices must st ate whether the layoffs relate to the employer's use of AI or another tec hnological change END:VALARM END:VEVENT BEGIN:VEVENT UID:us|fedreg|2026-16371|2026-10-13@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20261013 DTEND;VALUE=DATE:20261014 SUMMARY:United States — federal: NIST asks how the National Vulnerabilit y Database should be rebuilt for an AI-driven security landscape DESCRIPTION:NIST asks how the National Vulnerability Database should be re built for an AI-driven security landscape\nCommerce Department — Notice \nNIST asks how the National Vulnerability Database should be rebuilt for an AI-driven security landscape - which parts of the vulnerability lifec ycle to hand to AI automation\, where human review must stay\, how to mak e AI-driven risk prioritisation auditable\, and what safeguards are neede d against bad AI-generated fixes.\nApplies to: any interested party\nSour ce says: "comments_close_on: 2026-10-13 (Federal Register API)"\nSource: https://www.federalregister.gov/documents/2026/08/12/2026-16371/request-f or-information-rfi-on-modernizing-the-national-vulnerability-database-in- the-age-of\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-cal endar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Comments close TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: NIST asks how the National Vulnerab ility Database should be rebuilt for an AI-driven security landscape END:VALARM END:VEVENT BEGIN:VEVENT UID:us|fedreg|2026-17163|2026-10-20@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20261020 DTEND;VALUE=DATE:20261021 SUMMARY:United States — federal: CFTC asks how exchanges should design a nd list futures/swaps whose underlying commodity is AI compute capacity DESCRIPTION:CFTC asks how exchanges should design and list futures/swaps w hose underlying commodity is AI compute capacity\nCommodity Futures Tradi ng Commission — Proposed Rule\nCFTC asks how exchanges should design an d list futures/swaps whose underlying commodity is AI compute capacity - how to measure the size\, liquidity and price transparency of the compute cash market\, whether a contract may settle to an index the CFTC cannot verify\, and what standards fungible 'compute' delivery would require.\nA pplies to: any interested party\nSource says: "comments_close_on: 2026-10 -20 (Federal Register API)"\nSource: https://www.federalregister.gov/docu ments/2026/08/21/2026-17163/request-for-comment-on-the-listing-of-compute -derivatives-contracts\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-co mpliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Comments close TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — federal: CFTC asks how exchanges should desi gn and list futures/swaps whose underlying commodity is AI compute capaci ty END:VALARM END:VEVENT BEGIN:VEVENT UID:cl|ley-n-21-719-regula-la-protecci-n-y-el-tratamiento-de-los-datos-per son|automated-individual-decisions-and-profiling-data-subjects-g|2026-12- 01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20261201 DTEND;VALUE=DATE:20261202 SUMMARY:Chile: Automated individual decisions and profiling: data subjects gain the right to object and not to be subject to decisions based on aut omated processing that produce legal effects or significantly affect them \; controllers must in all cases guarantee information and transparency\, an explanation\, human intervention\, the right to express a point of vi ew and to request review of the decision\, and must disclose the existenc e of automated decisions together with meaningful information on the logi c applied DESCRIPTION:Automated individual decisions and profiling: data subjects ga in the right to object and not to be subject to decisions based on automa ted processing that produce legal effects or significantly affect them\; controllers must in all cases guarantee information and transparency\, an explanation\, human intervention\, the right to express a point of view and to request review of the decision\, and must disclose the existence o f automated decisions together with meaningful information on the logic a pplied\nLey N. 21.719 — Regula la protección y el tratamiento de los d atos personales y crea la Agencia de Protección de Datos Personales (pro mulgada 25-NOV-2024\, publicada en el Diario Oficial 13-DIC-2024)\, en pa rticular el nuevo Artículo 8° bis y el Artículo 14 letra l) que incorp ora a la Ley N. 19.628\nArt. 8° bis is Chile's GDPR-Art.22 analogue and is the operative AI/ADM hook. The exceptions are narrow (necessary for th e conclusion or performance of a contract\; prior express consent under a rt. 12\; where the law so provides with safeguards) and even inside those exceptions the safeguard duties still apply. Art. 14(l) adds a standing transparency duty to disclose 'la existencia de decisiones automatizadas\ , incluida la elaboración de perfiles' with 'información significativa sobre la lógica aplicada\, así como las consecuencias previstas de dich o tratamiento para el titular'. Profiling is defined in the new letter w) : 'toda forma de tratamiento automatizado de datos personales que consist a en utilizar esos datos para evaluar\, analizar o predecir aspectos rela tivos al rendimiento profesional\, situación económica\, de salud\, pre ferencias personales\, intereses\, fiabilidad\, comportamiento\, ubicaci ón o movimientos de una persona natural'. The law also creates the Agenc ia de Protección de Datos Personales as enforcement authority.\nApplies to: controller\, deployer\, processor\nWho: no size threshold\nPenalty: E nforced by the newly created Agencia de Protección de Datos Personales t hrough an administrative sanctioning regime graded by seriousness of infr ingement\, with fines in UTM and a Registro Nacional de Sanciones y Cumpl imiento\; the Agency also holds inspection and instruction powers.\nSourc e says: "Entra en vigencia el 01-DIC-2026 — LEY 21719 REGULA LA PROTECC IÓN Y EL TRATAMIENTO DE LOS DATOS PERSONALES Y CREA LA AGENCIA DE PROTEC CIÓN DE DATOS PERSONALES. Promulgación: 25-NOV-2024. Publicación: 13-D IC-2024. Versión: Con Vigencia Diferida por Fecha - 01-DIC-2026."\nSourc e: https://www.bcn.cl/leychile/navegar?idNorma=1209272\nVerified: Aug 30\ , 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Law applies TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Chile: Automated individual decisions and profiling: data subj ects gain the right to object and not to be subject to decisions based on automated processing that produce legal effects or significantly affect them\; controllers must in all cases guarantee information and transparen cy\, an explanation\, human intervention\, the right to express a point o f view and to request review of the decision\, and must disclose the exis tence of automated decisions together with meaningful information on the logic applied END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689|four-month-transitional-period-en ds-for-marking-generative-a|2026-12-02@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20261202 DTEND;VALUE=DATE:20261203 SUMMARY:European Union: Four-month transitional period ends for marking ge nerative AI systems already on the market before 2 August 2026 DESCRIPTION:Four-month transitional period ends for marking generative AI systems already on the market before 2 August 2026\nEU AI Act (Regulation (EU) 2024/1689)\nRegulation (EU) 2026/1744 gave providers whose generati ve systems were already placed on the market a four-month grace period on the Article 50 marking duty. After it\, no carve-out remains.\nApplies t o: provider\nPenalty: Up to EUR 15 000 000 or 3% of worldwide annual turn over\nSource says: "Providers of AI systems\, including general-purpose A I systems\, generating synthetic audio\, image\, video or text content\, that have been placed on the market before 2 August 2026 shall take the n ecessary steps in order to comply with Article 50(2) by 2 December 2026." \nSource: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:L_2 02601744\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calen dar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Four-month transitional period ends for markin g generative AI systems already on the market before 2 August 2026 END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|product-liability-directive-directive-eu-2024-2853|new-product-liab ility-regime-applies-to-software-and-ai-syst|2026-12-09@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20261209 DTEND;VALUE=DATE:20261210 SUMMARY:European Union: New product liability regime applies to software a nd AI systems placed on the market DESCRIPTION:New product liability regime applies to software and AI system s placed on the market\nProduct Liability Directive (Directive (EU) 2024/ 2853)\nSoftware — explicitly including AI systems\, whether embedded\, networked or delivered as SaaS — becomes a 'product' for strict liabili ty. AI system providers are treated as manufacturers. Applies to products placed on the market or put into service after 9 December 2026.\nApplies to: provider\, manufacturer\, importer\nPenalty: Strict civil liability for damage caused by defective products\; no financial ceiling\nSource sa ys: "Article 2(1): 'This Directive shall apply to products placed on the market or put into service after 9 December 2026.' Article 4(1): ‘produ ct’ means all movables\, even if integrated into\, or inter-connected w ith\, another movable or an immovable\; it includes electricity\, digital manufacturing files\, raw materials and software'. Article 22(1): 'Membe r States shall bring into force the laws\, regulations and administrative provisions necessary to comply with this Directive by 9 December 2026.'" \nSource: https://eur-lex.europa.eu/eli/dir/2024/2853/oj/eng\nVerified: A ug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Law applies TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: New product liability regime applies to softwa re and AI systems placed on the market END:VALARM END:VEVENT BEGIN:VEVENT UID:au|privacy-act-1988-cth-as-amended-by-the-privacy-and-other-legislatio n-a|privacy-policies-must-disclose-automated-decision-making-tha|2026-12- 10@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20261210 DTEND;VALUE=DATE:20261211 SUMMARY:Australia: Privacy policies must disclose automated decision-makin g that significantly affects people DESCRIPTION:Privacy policies must disclose automated decision-making that significantly affects people\nPrivacy Act 1988 (Cth)\, as amended by the Privacy and Other Legislation Amendment Act 2024\nA disclosure duty\, not a ban: APP entities must state in their privacy policy which personal in formation is used by computer programs to make decisions that significant ly affect individuals\, and what kinds of decisions those are. Reaches or dinary businesses\, not only AI companies\; human oversight does not exem pt.\nApplies to: deployer\, business\nWho: APP entities under the Privacy Act\nPenalty: Low-tier civil penalty regime under s 13K(1)(b)(iia) — i nfringement and compliance notices\, up to roughly AUD 330\,000 for a bod y corporate\nSource says: "The ADM obligation commences on 10 December 20 26. ... The ADM obligation will enhance the right to privacy by introduci ng requirements that entities must include information in privacy policie s about the kinds of personal information used in\, and types of decision s made by\, computer programs that use personal information to make decis ions that could reasonably be expected to significantly affect the rights or interests of an individual. (OAIC\, Automated Decision-Making Transpa rency Obligation (APP 1) Issues Paper\, 18 May 2026\, Executive summary). OAIC consultation page wording: "From 10 December 2026\, APP entities th at use personal information in ADM with the potential to affect rights or interests will be required to provide information in their privacy polic ies about the kinds of personal information used and the kinds of decisio ns made using ADM." Statutory text (Attachment A\, new APP 1.7): "Without limiting subclause 1.3\, the APP privacy policy of an APP entity must co ntain the information covered by subclause 1.8 if: (a) the entity has arr anged for a computer program to make\, or do a thing that is substantiall y and directly related to making\, a decision\; and (b) the decision coul d reasonably be expected to significantly affect the rights or interests of an individual\; and (c) personal information about the individual is u sed in the operation of the computer program to make the decision or do t he thing that is substantially and directly related to making the decisio n.""\nSource: https://www.oaic.gov.au/engage-with-us/consultations/consul tation-on-guidance-for-transparency-in-automated-decision-making\nVerifie d: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Australia: Privacy policies must disclose automated decision-m aking that significantly affects people END:VALARM END:VEVENT BEGIN:VEVENT UID:uk|crime-and-policing-act-2026-s-249-duty-to-make-progress-report-in-f orc|secretary-of-state-must-lay-before-parliament-a-progress-rep|2026-12- 31@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20261231 DTEND;VALUE=DATE:20270101 SUMMARY:United Kingdom: Secretary of State must lay before Parliament a pr ogress report on making regulations to bring illegal AI-generated content and AI services into the Online Safety Act 2023 DESCRIPTION:Secretary of State must lay before Parliament a progress repor t on making regulations to bring illegal AI-generated content and AI serv ices into the Online Safety Act 2023\nCrime and Policing Act 2026\, s.249 (duty to make progress report)\, in force at Royal Assent per s.255(2)(k )\nSection 248 of the Crime and Policing Act 2026 ('Power to amend Online Safety Act 2023: AI') inserts a new s.216A into the Online Safety Act 20 23\, letting the Secretary of State amend that Act by regulations to addr ess risks from illegal AI-generated content and the use of AI services to commit or facilitate priority offences. An 'AI service' is defined as an internet service capable of generating AI-generated content\, no matter what proportion of content on the service is AI-generated - the route by which standalone AI chatbots\, which Ofcom has said fall outside Part 3 w hen output is not user-generated or search content\, can be brought into the illegal content duties. Section 249 backs that power with a hard repo rting deadline: the Secretary of State must lay a progress report by 31 D ecember 2026 unless draft regulations have been laid before Parliament by then.\nApplies to: Secretary of State (duty holder)\, providers of AI se rvices and chatbots (prospective duty holders under the resulting regulat ions)\nWho: n/a\nPenalty: Parliamentary duty on the Secretary of State\; no financial penalty. The regulations it anticipates would carry full Onl ine Safety Act enforcement against AI services.\nSource says: "(1) The Se cretary of State must\, no later than 31 December 2026\, lay before Parli ament a report about the progress that has been made towards making regul ations under section 216A of the Online Safety Act 2023 (power to amend A ct in relation to illegal AI-generated content). (2) Subsection (1) does not apply if a draft of a statutory instrument containing regulations und er that section is laid before Parliament before 31 December 2026."\nSour ce: https://www.legislation.gov.uk/ukpga/2026/20/section/249\nVerified: A ug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Reporting TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United Kingdom: Secretary of State must lay before Parliament a progress report on making regulations to bring illegal AI-generated con tent and AI services into the Online Safety Act 2023 END:VALARM END:VEVENT BEGIN:VEVENT UID:us-ca|ccpa-regulations-on-automated-decisionmaking-technology-cppa-eff -1-jan|businesses-using-admt-for-significant-decisions-must-comply|2027-0 1-01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270101 DTEND;VALUE=DATE:20270102 SUMMARY:United States — California: Businesses using ADMT for significan t decisions must comply in full: pre-use notice\, opt-out and access righ ts DESCRIPTION:Businesses using ADMT for significant decisions must comply in full: pre-use notice\, opt-out and access rights\nCCPA regulations on au tomated decisionmaking technology (CPPA\, eff. 1 Jan 2026)\nSignificant d ecisions include financial or lending services\, housing\, education\, em ployment or independent contracting opportunities and healthcare.\nApplie s to: business\, deployer\nPenalty: CCPA administrative fines up to $2\,6 63 per violation\, $7\,988 per intentional violation (CPPA enforcement)\n Source says: "CCPA Regulations (11 CCR) § 7200(b): “A business that us es ADMT for a significant decision prior to January 1\, 2027\, must be in compliance with the requirements of this Article no later than January 1 \, 2027. A business that uses ADMT on or after January 1\, 2027\, must be in compliance with the requirements of this Article any time it is using ADMT for a significant decision.”"\nSource: https://cppa.ca.gov/regula tions/pdf/ccpa_updates_cyber_risk_admt_appr_text.pdf\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — California: Businesses using ADMT for signif icant decisions must comply in full: pre-use notice\, opt-out and access rights END:VALARM END:VEVENT BEGIN:VEVENT UID:us-ca|california-ai-transparency-act-sb-942-as-amended-by-ab-853|large -online-platforms-must-detect-and-surface-content-prove|2027-01-01@agencc y.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270101 DTEND;VALUE=DATE:20270102 SUMMARY:United States — California: Large online platforms must detect a nd surface content provenance data DESCRIPTION:Large online platforms must detect and surface content provena nce data\nCalifornia AI Transparency Act (SB 942 as amended by AB 853)\nP latforms above 2 million monthly users must detect provenance data in dis tributed content\, expose it in the user interface\, and let users inspec t whether content was AI-generated or captured by a device.\nApplies to: platform\nWho: over 2 million monthly users\nSource says: "Bus. & Prof. C ode § 22757.3.1 (added by AB 853\, Ch. 674\, Stats. 2025)\, operative-da te sentence: “This section shall become operative on January 1\, 2027. ” The duty it dates\, in the section's own words\, is to “detect whet her any provenance data that is compliant with widely adopted specificati ons adopted by an established standards-setting body is embedded into or attached to content.”"\nSource: https://leginfo.legislature.ca.gov/face s/billTextClient.xhtml?bill_id=202520260AB853\nVerified: Aug 30\, 2026\nh ttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — California: Large online platforms must dete ct and surface content provenance data END:VALARM END:VEVENT BEGIN:VEVENT UID:us-co|colorado-sb-26-189-automated-decision-making-technology|develope rs-and-deployers-of-admt-used-in-consequential-decis|2027-01-01@agenccy.a i DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270101 DTEND;VALUE=DATE:20270102 SUMMARY:United States — Colorado: Developers and deployers of ADMT used in consequential decisions must give notice\, documentation and post-deci sion disclosures DESCRIPTION:Developers and deployers of ADMT used in consequential decisio ns must give notice\, documentation and post-decision disclosures\nColora do SB 26-189\, Automated Decision-Making Technology\nReplaces the repeale d Colorado AI Act (SB 24-205). Developers must hand deployers technical d ocumentation on intended uses\, training-data categories\, known limitati ons and human-review instructions\; both sides keep compliance records fo r three years.\nApplies to: developer\, deployer\nPenalty: Enforcement ce ntralised with the Colorado Attorney General\; fault-based apportionment of discrimination liability\nSource says: "From the enrolled Final Act: " SECTION 5. Effective date - applicability. (1) Except as otherwise provid ed in subsection (2) of this section\, this act takes effect January 1\, 2027. ... (3) This act applies to consequential decisions made on or afte r January 1\, 2027." Operative duty: "6-1-1702. Developer responsibilitie s - documentation. (1) ON AND AFTER JANUARY 1\, 2027\, A DEVELOPER SHALL MAKE AVAILABLE TO EACH DEPLOYER OF A COVERED ADMT DEVELOPED BY THE DEVELO PER\, IN A FORM AND MANNER THAT IS REASONABLY UNDERSTANDABLE TO A DEPLOYE R AND THAT PROTECTS TRADE SECRETS...". Deployer notice: "[A DEPLOYER] SHA LL PROVIDE A CLEAR AND CONSPICUOUS NOTICE TO A CONSUMER THAT THE DEPLOYER USED OR WILL USE A COVERED ADMT IN A CONSEQUENTIAL DECISION AFFECTING TH E CONSUMER AND INSTRUCTIONS REGARDING HOW THE CONSUMER MAY OBTAIN THE ADD ITIONAL INFORMATION DESCRIBED IN THIS SECTION." AG rulemaking (BOTH provi sions): "(b) ON OR BEFORE JANUARY 1\, 2027\, THE ATTORNEY GENERAL SHALL A DOPT RULES TO CLARIFY AND IMPLEMENT THE POST-ADVERSE OUTCOME DISCLOSURE R EQUIREMENTS SET FORTH IN SUBSECTION (3) OF THIS SECTION" and "(3) ON OR B EFORE JANUARY 1\, 2027\, THE ATTORNEY GENERAL SHALL ADOPT RULES TO CLARIF Y AND IMPLEMENT THE REQUIREMENTS OF THIS SECTION." Fault apportionment: " FAULT SHALL BE ALLOCATED AMONG DEPLOYERS AND DEVELOPERS BASED ON THEIR RE LATIVE FAULT FOR THE VIOLATION.""\nSource: https://leg.colorado.gov/bills /sb26-189\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-cale ndar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — Colorado: Developers and deployers of ADMT u sed in consequential decisions must give notice\, documentation and post- decision disclosures END:VALARM END:VEVENT BEGIN:VEVENT UID:us-co|colorado-sb-26-189-automated-decision-making-technology|ag-must- adopt-rules|2027-01-01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270101 DTEND;VALUE=DATE:20270102 SUMMARY:United States — Colorado: Attorney General must adopt implementi ng rules — on the same day compliance begins DESCRIPTION:Attorney General must adopt implementing rules — on the same day compliance begins\nColorado SB 26-189\, Automated Decision-Making Te chnology\nTwo mandatory rulemaking duties\, 6-1-1704(4)(b) and 6-1-1705(3 )\, both worded 'on or before January 1\, 2027\, the Attorney General sha ll adopt rules'. Deployers therefore may not see final rules before their own obligations start.\nApplies to: Colorado Attorney General\nSource sa ys: "ON OR BEFORE JANUARY 1\, 2027\, THE ATTORNEY GENERAL SHALL ADOPT RUL ES"\nSource: https://leg.colorado.gov/bills/sb26-189\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Regulator milestone TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — Colorado: Attorney General must adopt implem enting rules — on the same day compliance begins END:VALARM END:VEVENT BEGIN:VEVENT UID:us-ny|responsible-ai-safety-and-education-act-raise-act|frontier-model -developers-must-publish-a-safety-framework-fi|2027-01-01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270101 DTEND;VALUE=DATE:20270102 SUMMARY:United States — New York: Frontier model developers must publish a safety framework\, file disclosures and report safety incidents within 72 hours DESCRIPTION:Frontier model developers must publish a safety framework\, fi le disclosures and report safety incidents within 72 hours\nResponsible A I Safety and Education Act (RAISE Act)\nBase law is Chapter 699 of 2025 ( S6953-B\, signed 19 December 2025)\, overhauled by chapter amendment S882 8/A9449 signed 27 March 2026\, which struck the original commencement and wrote in 1 January 2027. The amendment also moved oversight to a new off ice inside NYDFS.\nApplies to: developer\nWho: over $500M annual revenue\ nPenalty: Civil penalties up to $1\,000\,000 (first violation) and $3\,00 0\,000 (subsequent)\, enforced through NYDFS — reduced from $10M/$30M b y the March 2026 amendment\nSource says: "Effective-date clause as amende d (S8828 sec. 3\, showing the bracketed repeal and new date): "This act s hall take effect [on the ninetieth day after it shall have become a law] JANUARY 1\, 2027." Operative duties in the consolidated text: "A LARGE FR ONTIER DEVELOPER SHALL WRITE\, IMPLEMENT\, COMPLY WITH\, AND CLEARLY AND CONSPICUOUSLY PUBLISH ON ITS INTERNET WEBSITE A FRONTIER AI FRAMEWORK"\; "A LARGE FRONTIER DEVELOPER SHALL TRANSMIT TO THE OFFICE A SUMMARY OF ANY ASSESSMENT OF CATASTROPHIC RISK RESULTING FROM INTERNAL USE OF ITS FRONT IER MODELS EVERY THREE MONTHS"\; "A FRONTIER DEVELOPER SHALL REPORT ANY C RITICAL SAFETY INCIDENT PERTAINING TO ONE OR MORE OF ITS FRONTIER MODELS TO THE OFFICE WITHIN SEVENTY-TWO HOURS". Threshold definition: "A FRONTIE R DEVELOPER THAT TOGETHER WITH ITS AFFILIATES COLLECTIVELY HAD ANNUAL GRO SS REVENUES IN EXCESS OF FIVE HUNDRED MILLION DOLLARS IN THE PRECEDING CA LENDAR YEAR.""\nSource: https://www.nysenate.gov/legislation/bills/2025/S 8828\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — New York: Frontier model developers must pub lish a safety framework\, file disclosures and report safety incidents wi thin 72 hours END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|machinery-regulation-regulation-eu-2023-1230|new-machinery-regime-a pplies-covering-digital-and-ai-driven|2027-01-20@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270120 DTEND;VALUE=DATE:20270121 SUMMARY:European Union: New machinery regime applies\, covering digital an d AI-driven risks for the first time DESCRIPTION:New machinery regime applies\, covering digital and AI-driven risks for the first time\nMachinery Regulation (Regulation (EU) 2023/1230 )\nReplaces the Machinery Directive. Safety assessment must account for A I-based control functions and human-robot collaboration\; machinery with AI safety components sits in the higher conformity-assessment category.\n Applies to: manufacturer\, importer\nPenalty: National penalties under Ar ticle 50 — effective\, proportionate and dissuasive\, and may include c riminal penalties for serious infringements\nSource says: "Article 54 'En try into force and application'\, second paragraph\, AS CORRECTED: 'It sh all apply from 20 January 2027.' The as-published OJ L 165 text reads '14 January 2027'\; the Corrigendum in OJ L 169\, 4.7.2023\, p. 35 states: ' 10. On page 39\, Article 54\, second paragraph: for: ‘14 January 2027 ’\, read: ‘20 January 2027’.' Article 54\, first paragraph (unchang ed): 'This Regulation shall enter into force on the twentieth day followi ng that of its publication in the Official Journal of the European Union. '"\nSource: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELE X%3A32023R1230R%2801%29\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-c ompliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Law applies TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: New machinery regime applies\, covering digita l and AI-driven risks for the first time END:VALARM END:VEVENT BEGIN:VEVENT UID:vn|prime-minister-s-decision-no-33-2026-qd-ttg-list-of-high-risk-ai-sy ste|legacy-high-risk-ai-systems-outside-health-education-finance|2027-03- 01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270301 DTEND;VALUE=DATE:20270302 SUMMARY:Vietnam: Legacy high-risk AI systems outside health/education/fina nce that were already in operation before 15 August 2026 must have comple ted their compliance obligations under the AI Law DESCRIPTION:Legacy high-risk AI systems outside health/education/finance t hat were already in operation before 15 August 2026 must have completed t heir compliance obligations under the AI Law\nPrime Minister's Decision N o. 33/2026/QD-TTg (List of high-risk AI systems) — compliance roadmap\, read with Law on AI No. 134/2025/QH15 Art. 35\nProviders (nhà cung cấ p) and deployers (bên triển khai) of AI systems on the high-risk list that were already deployed before the Decision took effect get a transiti on window. Systems may keep operating during the window\, except where th e competent state authority determines a system risks serious harm and or ders suspension or termination.\nApplies to: provider\, deployer\nWho: al l\nPenalty: Competent state authority may require suspension or terminati on of the system during the transition window if it poses a risk of serio us harm\nSource says: "Trước ngày 01/3/2027 đối với các hệ th ống trí tuệ nhân tạo thuộc các lĩnh vực còn lại trong Da nh mục. // 'Before 1 March 2027 for the artificial intelligence systems in the remaining fields on the List.'"\nSource: https://mst.gov.vn/46-he -thong-ai-duoc-xep-vao-nhom-rui-ro-cao-phai-quan-ly-nghiem-ngat-197260703 152945179.htm\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance- calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Vietnam: Legacy high-risk AI systems outside health/education/ finance that were already in operation before 15 August 2026 must have co mpleted their compliance obligations under the AI Law END:VALARM END:VEVENT BEGIN:VEVENT UID:ca|osfi-guideline-e-23-model-risk-management-2027-office-of-the-superi nte|federally-regulated-financial-institutions-must-have-enterpr|2027-05- 01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270501 DTEND;VALUE=DATE:20270502 SUMMARY:Canada — federal: Federally regulated financial institutions mus t have enterprise-wide\, risk-based model risk management covering ALL mo dels including artificial intelligence and machine learning models — mo del inventory\, risk tiering\, governance and accountability across the f ull model lifecycle from development through deployment\, monitoring and decommissioning DESCRIPTION:Federally regulated financial institutions must have enterpris e-wide\, risk-based model risk management covering ALL models including a rtificial intelligence and machine learning models — model inventory\, risk tiering\, governance and accountability across the full model lifecy cle from development through deployment\, monitoring and decommissioning\ nOSFI Guideline E-23 — Model Risk Management (2027)\, Office of the Sup erintendent of Financial Institutions (final version published 11 Septemb er 2025)\nE-23 is principles-based and organised around three pillars — model risk understood and managed enterprise-wide\; a risk-based approac h built on model inventory\, risk tiering and assurance\; and governance across the full model lifecycle. The 2027 version materially widens the 2 017 original: scope expands from deposit-taking institutions to ALL feder ally regulated financial institutions including insurers and foreign bran ches\, and the definition of 'model' is broadened so AI/ML systems fall s quarely inside\, with added context for AI/ML model risk. Policies\, proc edures and controls must be proportionate to the institution's size\, ris k profile\, complexity of operations and interconnectedness in the financ ial system.\nApplies to: deployer\, operator\, risk owner\nWho: all feder ally regulated financial institutions\, including foreign bank branches a nd foreign insurance company branches\; expectations are proportionate to size\, risk profile\, complexity and interconnectedness\nPenalty: OSFI g uidelines are supervisory expectations rather than statutory offences: no n-compliance is addressed through the supervisory process — findings an d recommendations\, elevated risk ratings and intervention staging\, and where warranted directions of compliance and other supervisory powers und er the Bank Act\, Insurance Companies Act and Trust and Loan Companies Ac t. No administrative monetary penalty attaches to the guideline itself.\n Source says: "Effective date: May 1\, 2027"\nSource: https://www.osfi-bsi f.gc.ca/en/guidance/guidance-library/guideline-e-23-model-risk-management -2027\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar / URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Canada — federal: Federally regulated financial institutions must have enterprise-wide\, risk-based model risk management covering AL L models including artificial intelligence and machine learning models — model inventory\, risk tiering\, governance and accountability across the full model lifecycle from development through deployment\, monitorin g and decommissioning END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689|commission-compliance-guidelines- for-high-risk-classificatio|2027-08-01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270801 DTEND;VALUE=DATE:20270802 SUMMARY:European Union: Commission guidance due on practical implementatio n of Articles 8(2)\, 9(10) and 17(3) DESCRIPTION:Commission guidance due on practical implementation of Article s 8(2)\, 9(10) and 17(3)\nEU AI Act (Regulation (EU) 2024/1689)\nGuidance on avoiding duplication between the AI Act and Annex I Section A product legislation — it serves the Annex I cohort whose obligations land on 2 August 2028.\nApplies to: European Commission\nSource says: "the practic al implementation of Article 8(2)\, Article 9(10) and Article 17(3) in ac cordance with the principle of complementarity and proportionality\, with a view to ensuring consistency\, avoiding duplication and minimising add itional burdens when complying with the requirements of this Regulation a nd the requirements of the Union harmonisation legislation listed in Sect ion A of Annex I\; such guidelines shall be published by 1 August 2027."\ nSource: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:L_20 2601744\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calend ar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Regulator milestone TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Commission guidance due on practical implement ation of Articles 8(2)\, 9(10) and 17(3) END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689|legacy-general-purpose-ai-models- placed-on-the-market-before|2027-08-02@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270802 DTEND;VALUE=DATE:20270803 SUMMARY:European Union: Legacy general-purpose AI models placed on the mar ket before 2 August 2025 must be brought into compliance DESCRIPTION:Legacy general-purpose AI models placed on the market before 2 August 2025 must be brought into compliance\nEU AI Act (Regulation (EU) 2024/1689)\nTwo-year grandfathering under Article 111(3) expires. Applies to GPAI models already on the market when the GPAI chapter started apply ing.\nApplies to: provider\nPenalty: Up to EUR 15 000 000 or 3% of worldw ide annual turnover (Art. 101)\nSource says: "Providers of general-purpos e AI models that have been placed on the market before 2 August 2025 shal l take the necessary steps in order to comply with the obligations laid d own in this Regulation by 2 August 2027."\nSource: https://ai-act-service -desk.ec.europa.eu/en/ai-act/article-111\nVerified: Aug 30\, 2026\nhttps: //agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Legacy general-purpose AI models placed on the market before 2 August 2025 must be brought into compliance END:VALARM END:VEVENT BEGIN:VEVENT UID:vn|prime-minister-s-decision-no-33-2026-qd-ttg-list-of-high-risk-ai-sy ste|legacy-high-risk-ai-systems-in-health-education-and-finance|2027-09-0 1@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270901 DTEND;VALUE=DATE:20270902 SUMMARY:Vietnam: Legacy high-risk AI systems in health\, education and fin ance that were already in operation before 15 August 2026 must have compl eted their compliance obligations under the AI Law DESCRIPTION:Legacy high-risk AI systems in health\, education and finance that were already in operation before 15 August 2026 must have completed their compliance obligations under the AI Law\nPrime Minister's Decision No. 33/2026/QD-TTg (List of high-risk AI systems) — compliance roadmap\ , read with Law on AI No. 134/2025/QH15 Art. 35\nLonger transition window (than other sectors) for AI systems already deployed in health care\, ed ucation and finance. Systems may keep operating during the window\, excep t where the competent state authority determines a system risks serious h arm and orders suspension or termination.\nApplies to: provider\, deploye r\nWho: all\nPenalty: Competent state authority may require suspension or termination of the system during the transition window if it poses a ris k of serious harm\nSource says: "Trước ngày 01/9/2027 đối với c ác hệ thống trí tuệ nhân tạo trong lĩnh vực y tế\, giáo dục và tài chính. // 'Before 1 September 2027 for artificial intelli gence systems in the health\, education and finance sectors.'"\nSource: h ttps://mst.gov.vn/46-he-thong-ai-duoc-xep-vao-nhom-rui-ro-cao-phai-quan-l y-nghiem-ngat-197260703152945179.htm\nVerified: Aug 30\, 2026\nhttps://ag enccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:Vietnam: Legacy high-risk AI systems in health\, education and finance that were already in operation before 15 August 2026 must have c ompleted their compliance obligations under the AI Law END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689|commission-post-market-monitoring -guidance-due|2027-09-02@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20270902 DTEND;VALUE=DATE:20270903 SUMMARY:European Union: Commission post-market monitoring guidance due DESCRIPTION:Commission post-market monitoring guidance due\nEU AI Act (Reg ulation (EU) 2024/1689)\nApplies to: European Commission\nSource says: "T he Commission\, taking utmost account of the opinion of the Board\, shall adopt guidance\, including a template\, on the post-market monitoring pl an by 2 September 2027."\nSource: https://eur-lex.europa.eu/legal-content /EN/TXT/HTML/?uri=OJ:L_202601744\nVerified: Aug 30\, 2026\nhttps://agencc y.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Regulator milestone TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Commission post-market monitoring guidance due END:VALARM END:VEVENT BEGIN:VEVENT UID:us-ct|connecticut-ai-responsibility-and-transparency-act-public-act-26 -15-sb|employers-using-automated-employment-decision-technology-mus|2027- 10-01@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20271001 DTEND;VALUE=DATE:20271002 SUMMARY:United States — Connecticut: Employers using automated employmen t decision technology must give detailed notices to applicants and employ ees DESCRIPTION:Employers using automated employment decision technology must give detailed notices to applicants and employees\nConnecticut AI Respons ibility and Transparency Act (Public Act 26-15\, SB 5)\nApplies to: emplo yer\, deployer\nSource says: "Closest verified wording: from 1 October 20 27\, deployers of automated employment-related decision technology must d isclose to applicants and employees that they are interacting with AEDT u nless it is obvious\, and where AEDT is used to make\, or its output is a substantial factor in making\, an employment-related decision\, must giv e written notice before the decision covering the fact of its use\, the p urposes and the employment decisions affected\, the trade name of the tec hnology\, the categories of personal data analysed and how they are asses sed\, and employer contact details\; for an adverse decision the notice m ust give "a high-level statement disclosing the principal reasons for the decision\, including how much and how the automated process output contr ibuted to the decision\, and the type and source of data." AEDT is define d as "any technology that processes personal data and uses computation to generate any output\, including predictions\, recommendations\, classifi cations\, rankings\, scores or other information\, that is a substantial factor used to make or materially influence an employment-related decisio n." The row's source_quote is a commentator's paraphrase\, not statutory text."\nSource: https://www.cga.ct.gov/2026/act/pa/pdf/2026PA-00015-R00SB -00005-PA.pdf\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance- calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — Connecticut: Employers using automated emplo yment decision technology must give detailed notices to applicants and em ployees END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689-as-amended-by-regulation-eu-2026- 174|high-risk-obligations-apply-to-stand-alone-annex-iii-systems|2027-12- 02@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20271202 DTEND;VALUE=DATE:20271203 SUMMARY:European Union: High-risk obligations apply to stand-alone Annex I II systems — deferred from 2 August 2026 DESCRIPTION:High-risk obligations apply to stand-alone Annex III systems — deferred from 2 August 2026\nEU AI Act (Regulation (EU) 2024/1689)\, as amended by Regulation (EU) 2026/1744\nRisk management\, data governanc e\, technical documentation\, logging\, human oversight\, accuracy and ro bustness\, conformity assessment and EU database registration for stand-a lone high-risk systems: employment\, credit\, education\, essential servi ces\, law enforcement\, migration\, justice\, biometrics.\nApplies to: pr ovider\, deployer\, importer\, distributor\nPenalty: Up to EUR 15 000 000 or 3% of worldwide annual turnover (Art. 99(4))\nSource says: "Chapter I II\, Sections 1\, 2\, and 3\, with the exception of Article 6(5)\, shall apply from: (i) 2 December 2027 as regards AI systems classified as high- risk pursuant to Article 6(2) and Annex III\; and (ii) 2 August 2028 as r egards AI systems classified as high-risk pursuant to Article 6(1) and An nex I\;"\nSource: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?ur i=OJ:L_202601744\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-complian ce-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: High-risk obligations apply to stand-alone Ann ex III systems — deferred from 2 August 2026 END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|cyber-resilience-act-regulation-eu-2024-2847|full-cra-regime-applie s-security-by-design-vulnerability-han|2027-12-11@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20271211 DTEND;VALUE=DATE:20271212 SUMMARY:European Union: Full CRA regime applies: security by design\, vuln erability handling\, CE marking for products with digital elements DESCRIPTION:Full CRA regime applies: security by design\, vulnerability ha ndling\, CE marking for products with digital elements\nCyber Resilience Act (Regulation (EU) 2024/2847)\nApplies to: manufacturer\, importer\, di stributor\nPenalty: Up to EUR 15 000 000 or 2.5% of total worldwide annua l turnover\, whichever is higher (Art. 64(2))\nSource says: "This Regulat ion shall apply from 11 December 2027. (Article 71(2)\, first subparagrap h)"\nSource: https://eur-lex.europa.eu/eli/reg/2024/2847/oj/eng\nVerified : Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Full CRA regime applies: security by design\, vulnerability handling\, CE marking for products with digital elements END:VALARM END:VEVENT BEGIN:VEVENT UID:us-ca|ccpa-regulations-on-risk-assessments-cppa-eff-1-jan-2026|risk-as sessments-must-be-completed-for-processing-activities|2027-12-31@agenccy. ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20271231 DTEND;VALUE=DATE:20280101 SUMMARY:United States — California: Risk assessments must be completed f or processing activities started before the regulations took effect and s till running DESCRIPTION:Risk assessments must be completed for processing activities s tarted before the regulations took effect and still running\nCCPA regulat ions on risk assessments (CPPA\, eff. 1 Jan 2026)\nApplies to: business\n Source says: "CCPA Regulations (11 CCR) § 7155(b): “For any processing activity identified in section 7150\, subsection (b)\, that the business initiated prior to [OAL to fill in the effective date of these regulatio ns] and that continues after [OAL to fill in the effective date of these regulations]\, the business must conduct\, and document as set forth in s ection 7152\, a risk assessment in accordance with the requirements of th is Article no later than December 31\, 2027. The business must comply wit h the submission requirements set forth in section 7157\, subsection (a)( 1).”"\nSource: https://cppa.ca.gov/regulations/pdf/ccpa_updates_cyber_r isk_admt_appr_text.pdf\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-co mpliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — California: Risk assessments must be complet ed for processing activities started before the regulations took effect a nd still running END:VALARM END:VEVENT BEGIN:VEVENT UID:us-ca|california-ai-transparency-act-sb-942-as-amended-by-ab-853|captu re-device-manufacturers-must-embed-latent-provenance-di|2028-01-01@agencc y.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20280101 DTEND;VALUE=DATE:20280102 SUMMARY:United States — California: Capture device manufacturers must em bed latent provenance disclosures by default DESCRIPTION:Capture device manufacturers must embed latent provenance disc losures by default\nCalifornia AI Transparency Act (SB 942 as amended by AB 853)\nManufacturer name\, device name and version\, and content creati on date/time embedded in captured content\, with a user option to include them.\nApplies to: manufacturer\nSource says: "Bus. & Prof. Code § 2275 7.3.3 (added by AB 853\, Ch. 674\, Stats. 2025)\, operative-date sentence : “This section shall become operative on January 1\, 2028.” The duty it dates reads “Embed latent disclosures in content captured by the de vice by default.” The Legislative Counsel's Digest states the scope: “This bill would require\, beginning January 1\, 2028\, a capture devic e manufacturer\, with respect to any capture device the capture device ma nufacturer first produced for sale in the state on or after January 1\, 2 028\, to\, among other things\, provide a user with the option to include a latent disclosure in content captured by the capture device that conve ys certain information\, including the name of the capture device manufac turer.”"\nSource: https://leginfo.legislature.ca.gov/faces/billTextClie nt.xhtml?bill_id=202520260AB853\nVerified: Aug 30\, 2026\nhttps://agenccy .ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — California: Capture device manufacturers mus t embed latent provenance disclosures by default END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689-as-amended-by-regulation-eu-2026- 174|notified-bodies-18-month-conformity-assessment-transition-en|2028-01- 28@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20280128 DTEND;VALUE=DATE:20280129 SUMMARY:European Union: Deadline for notified bodies to apply for designat ion under the 18-month transition DESCRIPTION:Deadline for notified bodies to apply for designation under th e 18-month transition\nEU AI Act (Regulation (EU) 2024/1689)\, as amended by Regulation (EU) 2026/1744\nApplies to: notified body\nSource says: "W ithout prejudice to Article 28\, such notified bodies which have been not ified under the Union harmonisation legislation in Section A of Annex I\, shall apply for designation in accordance with Section 4 of this Chapter by 28 January 2028."\nSource: https://eur-lex.europa.eu/legal-content/EN /TXT/HTML/?uri=OJ:L_202601744\nVerified: Aug 30\, 2026\nhttps://agenccy.a i/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Deadline for notified bodies to apply for desi gnation under the 18-month transition END:VALARM END:VEVENT BEGIN:VEVENT UID:us-ca|ccpa-regulations-on-risk-assessments-cppa-eff-1-jan-2026|first-r isk-assessment-submission-to-the-cppa-covering-assess|2028-04-01@agenccy. ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20280401 DTEND;VALUE=DATE:20280402 SUMMARY:United States — California: First risk-assessment submission to the CPPA — covering assessments conducted in 2026 and 2027 DESCRIPTION:First risk-assessment submission to the CPPA — covering asse ssments conducted in 2026 and 2027\nCCPA regulations on risk assessments (CPPA\, eff. 1 Jan 2026)\nApplies to: business\nSource says: "CCPA Regula tions (11 CCR) § 7157(a)(1): “For risk assessments conducted in 2026 a nd 2027\, the business must submit to the Agency the information required by subsection (b) no later than April 1\, 2028.”"\nSource: https://cpp a.ca.gov/regulations/pdf/ccpa_updates_cyber_risk_admt_appr_text.pdf\nVeri fied: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Reporting TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:United States — California: First risk-assessment submission to the CPPA — covering assessments conducted in 2026 and 2027 END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689-as-amended-by-regulation-eu-2026- 174|high-risk-obligations-apply-to-ai-embedded-in-regulated-prod|2028-08- 02@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20280802 DTEND;VALUE=DATE:20280803 SUMMARY:European Union: High-risk obligations apply to AI embedded in regu lated products (Annex I) — deferred from 2 August 2027 DESCRIPTION:High-risk obligations apply to AI embedded in regulated produc ts (Annex I) — deferred from 2 August 2027\nEU AI Act (Regulation (EU) 2024/1689)\, as amended by Regulation (EU) 2026/1744\nMachinery\, medical devices\, vehicles\, lifts\, toys\, radio equipment and the other Annex I product regimes. Corresponding machinery AI requirements land on the sa me date.\nApplies to: provider\, manufacturer\nPenalty: Up to EUR 15 000 000 or 3% of worldwide annual turnover\nSource says: "Chapter III\, Secti ons 1\, 2\, and 3\, with the exception of Article 6(5)\, shall apply from : ... (ii) 2 August 2028 as regards AI systems classified as high-risk pu rsuant to Article 6(1) and Annex I\;"\nSource: https://eur-lex.europa.eu/ legal-content/EN/TXT/HTML/?uri=OJ:L_202601744\nVerified: Aug 30\, 2026\nh ttps://agenccy.ai/ai-compliance-calendar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: High-risk obligations apply to AI embedded in regulated products (Annex I) — deferred from 2 August 2027 END:VALARM END:VEVENT BEGIN:VEVENT UID:eu|eu-ai-act-regulation-eu-2024-1689|large-scale-eu-it-systems-listed- in-annex-x-must-comply|2030-12-31@agenccy.ai DTSTAMP:20260831T023727Z DTSTART;VALUE=DATE:20301231 DTEND;VALUE=DATE:20310101 SUMMARY:European Union: Large-scale EU IT systems listed in Annex X must c omply DESCRIPTION:Large-scale EU IT systems listed in Annex X must comply\nEU AI Act (Regulation (EU) 2024/1689)\nSystems such as SIS\, VIS\, EES and ETI AS placed on the market before 2 August 2027 get until end-2030.\nApplies to: provider\, deployer\nSource says: "AI systems which are components o f the large-scale IT systems established by the legal acts listed in Anne x X that have been placed on the market or put into service before 2 Augu st 2027 shall be brought into compliance with this Regulation by 31 Decem ber 2030."\nSource: https://ai-act-service-desk.ec.europa.eu/en/ai-act/ar ticle-111\nVerified: Aug 30\, 2026\nhttps://agenccy.ai/ai-compliance-cale ndar/ URL:https://agenccy.ai/ai-compliance-calendar/ CATEGORIES:Compliance TRANSP:TRANSPARENT BEGIN:VALARM TRIGGER:-P7D ACTION:DISPLAY DESCRIPTION:European Union: Large-scale EU IT systems listed in Annex X mu st comply END:VALARM END:VEVENT END:VCALENDAR