Home / Reports / Global AI Regulation Tracker
Thank you for reading this post, don't forget to subscribe!
The single most consequential deadline in global AI law arrives on 2 August 2026, when the bulk of the EU AI Act — including obligations for high-risk systems, governance and the penalty regime — becomes applicable, exposing violators to fines of up to €35 million or 7% of worldwide annual turnover, whichever is higher (European Commission, Regulation (EU) 2024/1689, Article 99). This tracker is a July 2026 snapshot; every date and figure below is drawn from a named government, EU or regulatory source, listed at the end.
European Union: the phased rollout
The EU AI Act (Regulation (EU) 2024/1689) entered into force on 1 August 2024 and applies in stages. The ban on “unacceptable-risk” practices — including social scoring, untargeted facial-image scraping and certain manipulative or biometric systems — took effect on 2 February 2025. Obligations for general-purpose AI (GPAI) models and the governance architecture began on 2 August 2025. The core of the regime, covering high-risk systems, market surveillance, sanctions and most remaining provisions, applies from 2 August 2026, with full applicability reached on 2 August 2027 (European Commission, AI Act implementation timeline).
The penalty ceiling is the highest in the world: up to €35 million or 7% of total worldwide annual turnover for breaching the Article 5 prohibitions, with lower tiers for other violations (EU AI Act, Article 99). Note: following a political agreement reached on 7 May 2026 as part of the EU’s “digital omnibus” package, the Commission has signalled that application of rules for certain high-risk Annex III use cases (biometrics, critical infrastructure, employment, migration and others) would move to 2 December 2027. Readers should treat this specific deferral as reported-but-evolving, since the amending instrument was still moving through the legislative process at the time of writing.
United States: federal action and a preemption fight
On 2 June 2026, President Trump signed Executive Order 14409, “Promoting Advanced Artificial Intelligence Innovation and Security,” published in the Federal Register on 5 June 2026. The order is deregulatory and security-focused: it directs federal cyber-defense upgrades, establishes a voluntary framework for “covered frontier models,” and prioritizes criminal enforcement against AI-enabled cyber-crime. Critically, it explicitly declines to impose mandatory licensing, preclearance or permitting on AI developers (White House; Federal Register, doc. 2026-11415).
Congress has not enacted a comprehensive federal AI statute. On 4 June 2026, Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a bipartisan discussion draft of the “Great American AI Act” — a draft for stakeholder feedback, not an introduced bill. It proposes frontier-AI transparency, safety-incident reporting and whistleblower protections, alongside a contested three-year preemption of state AI laws (Rep. Obernolte, press release; TechPolicy.Press; Roll Call).
State law therefore remains the operative layer. Texas’s Responsible AI Governance Act (TRAIGA), signed 22 June 2025, took effect 1 January 2026. Colorado repealed and replaced its 2024 AI Act (SB 24-205) via SB 26-189, signed 14 May 2026 and effective 1 January 2027, pivoting toward a lighter disclosure-and-transparency model (Colorado General Assembly; Littler; Perkins Coie).
United Kingdom: principles, not a statute
The UK has deliberately avoided an omnibus AI law. As of 2026, no comprehensive AI Act has passed Parliament. Instead, a pro-innovation, sector-led framework asks existing regulators — the ICO, FCA, MHRA, CMA and EHRC — to apply five cross-cutting principles: safety and robustness; transparency and explainability; fairness; accountability and governance; and contestability and redress (House of Commons Library briefing CBP-10003).
China: content-specific control
China regulates through targeted instruments rather than a single act. The Interim Measures for the Management of Generative AI Services took effect on 15 August 2023, and mandatory AI-content labeling rules took effect on 1 September 2025, requiring both explicit and implicit labels on synthetic media. Draft measures targeting “anthropomorphic”/human-like interactive AI were published in late 2025 (China Law Translate; White & Case AI Watch).
Rest of the world
Canada’s Artificial Intelligence and Data Act (AIDA), part of Bill C-27, died on the order paper when Parliament was prorogued in January 2025; federal AI governance now rests on the Treasury Board’s Directive on Automated Decision-Making, in force since 2019 (ISED Canada; Parliament of Canada LEGISinfo). Globally, the OECD.AI Policy Observatory tracks AI policies across more than 70 countries and territories plus the EU, and its Policy Navigator catalogues over 1,300 initiatives from 80+ jurisdictions and international bodies (OECD.AI).
Frequently asked questions
When does the EU AI Act start to bite?
Prohibitions applied from 2 February 2025 and GPAI rules from 2 August 2025, but the main body of obligations — high-risk systems, governance and penalties — applies from 2 August 2026, with full applicability on 2 August 2027 (European Commission).
What is the maximum EU AI Act fine?
Up to €35 million or 7% of total worldwide annual turnover, whichever is higher, for breaching the Article 5 prohibitions (EU AI Act, Article 99).
Does the United States have a federal AI law?
Not a comprehensive one. Executive Order 14409 (June 2026) sets a voluntary, security-focused federal policy, and the bipartisan “Great American AI Act” exists only as a June 2026 discussion draft. Binding rules today are largely state-level, such as Texas’s TRAIGA (effective 1 January 2026).
Sources
- European Commission, “AI Act — Regulatory framework for AI” (implementation timeline), accessed 2026 — https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
- EU Artificial Intelligence Act, “Article 99: Penalties” — https://artificialintelligenceact.eu/article/99/
- The White House, “Promoting Advanced Artificial Intelligence Innovation and Security” (Executive Order 14409), 2 June 2026 — https://www.whitehouse.gov/presidential-actions/2026/06/promoting-advanced-artificial-intelligence-innovation-and-security/
- Federal Register, “Promoting Advanced Artificial Intelligence Innovation and Security,” 5 June 2026 (doc. 2026-11415) — https://www.federalregister.gov/documents/2026/06/05/2026-11415/promoting-advanced-artificial-intelligence-innovation-and-security
- Rep. Jay Obernolte, “Obernolte, Trahan release a discussion draft of the Great American AI Act,” 4 June 2026 — https://obernolte.house.gov/media/press-releases/obernolte-trahan-release-discussion-draft-great-american-ai-act
- Roll Call, “Bipartisan AI draft proposes three-year preemption of state laws,” 4 June 2026 — https://rollcall.com/2026/06/04/bipartisan-ai-draft-proposes-three-year-preemption-of-state-laws/
- Perkins Coie, “Texas Enacts Responsible Artificial Intelligence Governance Act,” 2025 — https://perkinscoie.com/insights/update/texas-enacts-responsible-artificial-intelligence-governance-act
- Littler, “Colorado Amends its Artificial Intelligence Law” (SB 26-189), May 2026 — https://www.littler.com/news-analysis/asap/colorado-amends-its-artificial-intelligence-law-substantially-reducing
- UK House of Commons Library, “AI regulation in the UK” (CBP-10003) — https://commonslibrary.parliament.uk/research-briefings/cbp-10003/
- China Law Translate, “Interim Measures for the Management of Generative AI Services” — https://www.chinalawtranslate.com/en/generative-ai-interim/
- ISED Canada, “The Artificial Intelligence and Data Act (AIDA)”; Parliament of Canada, Bill C-27 (44-1) — https://ised-isde.canada.ca/site/innovation-better-canada/en/artificial-intelligence-and-data-act ; https://www.parl.ca/legisinfo/en/bill/44-1/c-27
- OECD.AI Policy Observatory, “Live repository of AI strategies & policies” / Policy Navigator — https://oecd.ai/en/dashboards/national