{"id":915,"date":"2026-07-28T13:39:00","date_gmt":"2026-07-28T04:39:00","guid":{"rendered":"https:\/\/onepress.co.kr\/index.php\/briefing\/2026-07-28-eu-ai-omnibus-2026-en\/"},"modified":"2026-07-28T13:39:00","modified_gmt":"2026-07-28T04:39:00","slug":"2026-07-28-eu-ai-omnibus-2026-en","status":"publish","type":"briefing","link":"https:\/\/onepress.co.kr\/index.php\/briefing\/2026-07-28-eu-ai-omnibus-2026-en\/","title":{"rendered":"EU delays high-risk AI rules to 2027\u20132028, while sexual deepfake bans start in December 2026"},"content":{"rendered":"

2026-07-28 13:39 KST<\/strong><\/p>\n

The EU AI Omnibus entered into force on July 27, 2026.<\/p>\n

High-risk AI obligations move to 2027 or 2028 depending on the category, while new prohibitions on AI involving non-consensual intimate material and child sexual abuse material apply from December 2026.<\/p>\n

The three dates to separate first<\/h2>\n

Different parts of the same law now follow different clocks. Operators should classify the system and its market date rather than conclude that the entire AI Act was delayed.<\/p>\n

    \n
  1. December 2, 2026: new prohibited practices for intimate and child sexual abuse material, plus the Article 50(2) transition deadline for generative systems placed on the market before August 2.<\/li>\n
  2. December 2, 2027: relevant duties for stand-alone Annex III high-risk AI in areas such as education, employment, essential services and law enforcement.<\/li>\n
  3. August 2, 2028: relevant duties for high-risk AI embedded in Annex I products such as medical devices, machinery, toys and lifts.<\/li>\n<\/ol>\n

    Why the high-risk timetable moved<\/h2>\n

    The relevant high-risk rules had been due to apply on August 2, 2026. The EU found that delayed standards, common specifications, guidance and national authorities could create implementation costs without improving protection.<\/p>\n

    The new dates are fixed rather than conditional. The regulation also extends some simplified treatment to small mid-cap companies, expands regulatory sandboxes and gives the AI Office broader oversight powers.<\/p>\n

    \"Produced
    Produced explanatory image: it represents supervised high-risk AI testing in a regulatory sandbox and is not a real EU institution, company or testing facility.<\/figcaption><\/figure>\n

    What the new prohibitions target<\/h2>\n

    The law prohibits certain AI systems that realistically generate or manipulate an identifiable person\u2019s intimate parts or sexual activity without freely given, specific and explicit consent. It also covers generation or manipulation of child sexual abuse material.<\/p>\n

    Providers are covered when such output is intended or reasonably foreseeable and reproducible without adequate safeguards. Deployers are covered when they use a system for that purpose. The rule does not ban every nude, medical, virtual try-on or artistic image.<\/p>\n

    Generative AI marking was not delayed wholesale<\/h2>\n

    Providers of synthetic audio, image, video or text systems already on the market before August 2, 2026 receive four months, until December 2, to meet the machine-readable marking duty in Article 50(2).<\/p>\n

    That is not a new start date for every transparency duty. Disclosures about AI interaction, deepfakes and other deployer situations still require separate reading of Article 50 and its guidance.<\/p>\n

    What operators need to distinguish now<\/h2>\n

    An organization serving the EU market should document whether its system is stand-alone Annex III or product-based Annex I, whether a generative system predates August 2, whether synthetic outputs are marked, and whether safeguards prevent and correct prohibited misuse.<\/p>\n

    As checked on July 28, 2026, this is directly applicable EU law. Classification, exceptions, penalties and reach to providers outside the EU depend on the system, market and users; this briefing is not a substitute for legal advice.<\/p>\n

    Official primary sources<\/h2>\n

    European Commission: entry into force and summary of the AI Omnibus<\/a><\/p>\n

    EUR-Lex: full text and application dates of Regulation (EU) 2026\/1744<\/a><\/p>\n

    Council of the EU: final approval, scope and implementation timeline<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

    The EU AI Omnibus entered into force on July 27, 2026. High-risk AI duties move to December 2027 or August 2028, but new prohibitions covering non-consensual intimate material and child sexual abuse material begin in December 2026.<\/p>\n","protected":false},"featured_media":0,"template":"","meta":[],"class_list":["post-915","briefing","type-briefing","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/onepress.co.kr\/index.php\/wp-json\/wp\/v2\/briefing\/915","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/onepress.co.kr\/index.php\/wp-json\/wp\/v2\/briefing"}],"about":[{"href":"https:\/\/onepress.co.kr\/index.php\/wp-json\/wp\/v2\/types\/briefing"}],"wp:attachment":[{"href":"https:\/\/onepress.co.kr\/index.php\/wp-json\/wp\/v2\/media?parent=915"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}