AI Regulation and Policy (state at 2026-10-01)
Verification (2026-10-02)
The European Commission’s AI Act page and AI Act Service Desk timeline (primary) confirm: AI Omnibus in force 27 July 2026; prohibitions 1 to 8 since 2 Feb 2025 and a ninth (non-consensual intimate content and child abuse material, introduced by the Omnibus) from December 2026 (service desk: 2 Dec 2026, with a transition period for certain providers’ content-related obligations); GPAI obligations since 2 Aug 2025; general application and enforcement from 2 Aug 2026; sandboxes 2 Aug 2027; Annex III high-risk 2 Dec 2027; Annex I 2 Aug 2028. White House EO 14365 (11 Dec 2025) and Colorado SB26-189 (signed 14 May 2026, effective 1 Jan 2027) were confirmed on the official sites. The regulation number 2026/1744, adoption date and Official Journal date still rest on EUR-Lex listings seen via search and law-firm notes: the EUR-Lex text did not load. Quote article wording from the Official Journal before relying on it legally. Not legal advice.
EU AI Act (Regulation (EU) 2024/1689) after the Digital Omnibus
The “Digital Omnibus on AI”, Regulation (EU) 2026/1744, amends the AI Act (plus the EASA and Machinery regulations). Path: political agreement May 2026, European Parliament vote 2026-06-16, Council final approval 2026-06-29, adopted/signed 2026-07-08, published in the Official Journal (L series) 2026-07-24, in force 2026-07-27 (six days before the original 2026-08-02 high-risk date).
| Date | What applies |
|---|---|
| 2025-02-02 | First eight prohibited practices (social scoring, workplace emotion recognition, untargeted face scraping…) and the Art. 4 AI literacy duty (softened by the Omnibus, see below) |
| 2025-08-02 | General-purpose AI (GPAI) obligations: transparency, copyright policy, documentation; governance |
| 2026-08-02 | AI Office and Member States start enforcing; GPAI fines possible (up to 3% of turnover or EUR 15m per the Act; not rechecked in the primary text); Art. 50 transparency (label AI-generated content, deepfakes, AI interaction disclosure) |
| 2026-12-02 | New ninth prohibition: AI generating non-consensual intimate imagery and CSAM; marking/detection grace period ends for generative systems already on the market before 2026-08-02 |
| 2027-08-02 | Member State regulatory sandboxes (postponed); GPAI models placed on the market before 2025-08-02 must comply |
| 2027-12-02 | Annex III stand-alone high-risk systems (biometrics, critical infrastructure, education, employment, migration, etc.), moved from 2026-08-02 |
| 2028-08-02 | High-risk AI embedded in products under sector law (Annex I: lifts, toys, machinery…) |
AI literacy (Art. 4): retained but softened. Providers and deployers must “take measures to support the development of AI literacy” of staff (previously: ensure, to their best extent, a sufficient level); the text adds that this does not require guaranteeing any specific level for any individual, and tasks the Commission and Member States with supporting SMEs. (Wording per law-firm analyses (lawandtechnology.eu, Freshfields) of the final text; earlier tracker sources that said the duty was dropped or unchanged are superseded.)
Also changed: SME simplifications extended to small mid-caps; the AI Office gains supervision over AI systems built on a GPAI model from the same provider; registration is reinstated for self-assessed “not high-risk” systems.
What matters for a Belgian user
- Belgium has not (as of mid-2026 tracker sources; nothing newer found) enacted an AI Act implementing law. The 2025-2029 federal government agreement (per the regulatoryai.eu tracker; not verified against the agreement itself) names BIPT/IBPT as lead market-surveillance authority, with FPS Economy coordinating; the data protection authority (GBA/APD) covers personal-data aspects. Per the tracker, Belgium missed the 2025-08-02 authority-designation deadline; the BIPT role appears to be policy, not statute. Per the tracker, Belgium has no operational national AI sandbox yet (EU deadline now 2027-08-02 after the Omnibus); the FARI test hub in Brussels is described as a testbed (not verified).
- Using tools (ChatGPT, Claude, Copilot) as a deployer: softened AI literacy duty (Art. 4) and Art. 50 transparency if you publish AI-generated content or run chatbots. Whether building on GPAI models makes you a GPAI provider depends on the Act’s definitions and Commission guidance; not verified here, seek legal advice.
- GDPR still applies in parallel, see data-privacy-in-ai.
Other regimes (compact)
| Region | Status |
|---|---|
| US federal | No comprehensive AI statute. Executive Order 14365 of 2025-12-11 (White House) creates a DOJ AI Litigation Task Force to challenge state AI laws, tasks Commerce to list “onerous” state laws and FCC/FTC to start preemption proceedings. White House non-binding legislative recommendations 2026-03-20 urge preemption. A bipartisan House discussion draft (Obernolte, Trahan et al., June 2026) would preempt state model-development laws for 3 years; secondary reporting says it had stalled as of September 2026; on 2026-09-23 a bipartisan group of 26 state attorneys general urged Congress to regulate frontier AI without pre-empting state laws (secondary: regulatoryoversight.com), and a Senate frontier-AI bill was reported as still unreleased |
| US states | California SB 53 (frontier-model transparency; large developers must publish a frontier AI framework, penalties up to USD 1M per violation) effective 2026-01-01; New York RAISE Act signed 2025-12-19, effective 2027-01-01; Colorado’s AI Act was repealed and replaced by SB 26-189 (signed 2026-05-14, effective 2027-01-01, automated decision-making disclosure model) |
| UK | No AI Act. The May 2026 King’s Speech contained no AI bill; AI is handled via sector regulators (ICO, FCA, Ofcom) and sandboxes (AI Growth Lab via the Regulating for Growth Bill); AI Security Institute is not a regulator |
| South Korea | AI Basic Act signed 2025-01-21, in force 2026-01-22 (user notification, impact assessments for high-impact AI, domestic representative for foreign providers); a one-year grace period before fines (secondary sources) |
| Japan | AI Promotion Act enacted 2025-05-28, fully in force 2025-09-01: soft law, no penalties or prohibitions (secondary sources) |
| Council of Europe | Framework Convention on AI opened for signature 2024-09-05; not in force until five ratifications (see ethics-in-ai) |
| China | Labelling Measures for AI-generated content and mandatory standard GB 45438-2025 in force since 2025-09-01; amended Cybersecurity Law with AI provisions in force 2026-01-01 (approved 2025-10-28). |
What changed in 2026
- EU: AI Act high-risk dates postponed (Annex III to 2027-12-02, Annex I to 2028-08-02); Art. 4 literacy softened; new Art. 5 ban on non-consensual intimate imagery/CSAM generation; Art. 50 grace for existing systems shortened to 2026-12-02.
- US: Colorado rewrite; federal preemption push (EO 14365, stalled House draft).
- UK: no AI bill in the 2026 King’s Speech.
- China: Cybersecurity Law AI clauses in force.
Limits
The regulation number 2026/1744, adoption date and Official Journal date rest on the EUR-Lex listing and law-firm notes (the full text did not load); Art. 50 grace wording is per law-firm analyses (the Commission service desk confirms a December 2026 transition period); Belgian authority status is per the regulatoryai.eu tracker; California SB 53’s start date is confirmed by law-firm alerts and Wikipedia (2026-10-05); the stalled House draft is per secondary reporting (Nextgov, June 2026; TechPolicy.Press September 2026 via search); Korea and Japan rows are from law-firm and tracker pages, not the statutes. This note is not legal advice.
Sources (accessed 2026-10-02 unless noted)
- AI Act Service Desk timeline (primary): https://ai-act-service-desk.ec.europa.eu/en/ai-act/timeline/timeline-implementation-eu-ai-act
- Colorado SB26-189 (official): https://leg.colorado.gov/bills/sb26-189
- California SB 53 (official): https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53
- (accessed 2026-10-01, earlier sources below)
- European Commission, AI Act regulatory framework: https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai (primary)
- Freshfields, final Digital Omnibus on AI: https://www.freshfields.com/en/our-thinking/blogs/technology-quotient/eu-ai-act-unpacked-34-the-final-digital-omnibus-on-ai-key-amendments-to-the-a-102nber
- Regulation (EU) 2026/1744 overview: https://www.nicfab.eu/en/posts/digital-omnibus-ai-official-journal/ and https://lawandtechnology.eu/en/digital-omnibus-on-ai-official-journal-regulation-2026-1744/
- Council adoption 2026-06-29: https://www.licentium.io/post/eu-council-ai-act-omnibus-amendments-june-2026
- Belgium: https://www.regulatoryai.eu/ai-act-belgium/ and https://www.glacis.io/guide-eu-ai-act-belgium
- US: https://www.paulhastings.com/insights/client-alerts/president-trump-signs-executive-order-challenging-state-ai-laws ; https://www.gtlaw.com/en/insights/2026/5/colorado-repeals-and-replaces-the-colorado-ai-act
- Korea/Japan/NY RAISE/SB 53 (search results 2026-10-05): https://www.stimson.org/2026/south-koreas-ai-basic-act-seeking-balance-between-industry-innovation-and-social-risk/ ; https://www.ibanet.org/japan-emerging-framework-ai-legislation-guidelines ; https://www.mofo.com/resources/insights/260105-new-york-enacts-the-raise-act-regulating-frontier-ai-models ; https://en.wikipedia.org/wiki/Transparency_in_Frontier_Artificial_Intelligence_Act ; https://www.regulatoryoversight.com/2026/10/26-state-ags-urge-congress-to-regulate-frontier-ai-and-preserve-state-authority/
- UK: https://bratby.law/uk-ai-regulation-what-the-law-says/
- China: https://www.whitecase.com/insight-our-thinking/ai-watch-global-regulatory-tracker-china
- European Commission AI Omnibus statement (same page): confirms in force 2026-07-27
- EUR-Lex listing: https://eur-lex.europa.eu/eli/reg/2026/1744/oj/eng (via search result; full text not fetched)
- White & Case, AI Omnibus enters into force: https://www.whitecase.com/insight-alert/eu-ai-omnibus-enters-force-amending-ai-act (via search)
- Art. 4 wording: https://lawandtechnology.eu/en/ai-literacy-digital-omnibus-article-4-ai-act/
- White House EO 14365: https://www.whitehouse.gov/presidential-actions/2025/12/eliminating-state-law-obstruction-of-national-artificial-intelligence-policy/
- US preemption draft: https://www.nextgov.com/artificial-intelligence/2026/06/lawmakers-propose-ai-framework-would-preempt-state-laws-3-years/413975/
- Colorado SB 26-189: https://www.seyfarth.com/news-insights/colorado-enacts-artificial-intelligence-replacement-law.html
- UK King’s Speech: https://iapp.org/news/a/king-s-speech-signals-diffuse-uk-digital-policy-agenda-but-no-ai-bill
- China CSL: https://www.reedsmith.com/articles/china-approves-major-amendments-to-cybersecurity-law/ ; https://www.china-briefing.com/news/china-cybersecurity-law-amendment/