Most AI rules start as advice. A country publishes a strategy, then ethics guidance, and sometimes later a law.
We analyzed AI laws and policy in 24 jurisdictions across the Americas, Europe and Asia-Pacific, as they stood on 17 Oct 2026. For each one we traced how its rules took shape, then compared them to find the requirements they share.
Every jurisdiction here backs AI growth with a strategy, plan or funding. Only 8 of 24 have a national law written for AI. In 23 of 24, data protection law is the base layer.
01 · The board
Where each jurisdiction stands
Each jurisdiction appears once, ranked by the strongest rule it has. A country with both a law and a growth plan is shown under the law. "AI-specific law" here means a national statute written for AI, which is why the US sits under existing law even though several of its states have passed AI statutes of their own.
1
Strategy first
Plans and policy, no rules
2
Voluntary first
Frameworks and guidelines
7
Existing law applies
Sector rules, no national AI statute
6
Bills in play
Draft or proposed laws
8
AI-specific law
A national statute written for AI
02 · Timeline
Legal timelines
Asia-Pacific has the most laws in force in our review, 12, and half of them took effect in 2026. Europe's count is smaller because one EU act covers 27 member states, and its 2026 entries amend that act or enforce it at home, as Ireland's new law does. In the Americas, the binding items are Peru's AI law and US federal and state statutes, while Latin America's broader bills are still in their legislatures. Strategies keep arriving in every region, which points to a push to expand AI use alongside the rules.
Americas
30 dated
5 in force
Europe
11 dated
5 in force
Asia-Pacific
35 dated
12 in force
03 · Pace
New measures by year
1 in 2018: 1 Strategy or funding
0 in 2019
2 in 2020: 1 Advice, voluntary, 1 Strategy or funding
2 in 2021: 2 Strategy or funding
2 in 2022: 1 Law in force, 1 Strategy or funding
8 in 2023: 3 Law in force, 1 Bill or draft, 3 Advice, voluntary, 1 Strategy or funding
12 in 2024: 2 Law in force, 3 Bill or draft, 6 Advice, voluntary, 1 Strategy or funding
22 in 2025: 8 Law in force, 6 Bill or draft, 2 Advice, voluntary, 6 Strategy or funding
27 in 2026: 8 Law in force, 8 Bill or draft, 4 Advice, voluntary, 7 Strategy or funding
04 · Regions
Regions
New laws and bills in each region, listed by country.
Americas
Latin America is borrowing the EU's risk-based model. Brazil's Senate passed PL 2338/2023 and the bill now waits in the Chamber. Argentina has three bills and Chile has a draft. Colombia's government bill was archived in 2025, and its latest risk-based bill, PL 025/2026, is waiting for a first debate in the House. Peru is the only one here with an AI law in force, Law 31814, whose duties phase in by sector from September 2026. The US has no comprehensive federal AI statute. States fill the gap with laws such as California's SB 53, Illinois' HB 3773 and Colorado's SB 26-189, which applies from 1 Jan 2027. A December 2025 executive order set up a federal task force to challenge state AI laws, and a House discussion draft from June 2026, the Great American AI Act, would preempt some of them. Canada's AI and Data Act died, and a new privacy bill was tabled in June 2026.
- 2024ArgentinaBill 3003-D-2024, an EU AI Act style risk-based bill. Still under deliberation as of Jun 2026Bill or draft
- May 2025ArgentinaBill 2397-D-2025, a national AI oversight committeeBill or draft
- Aug 2025ArgentinaBill 4243-D-2025, personal data in AI systemsBill or draft
- 2023BrazilPL 2338/2023, comprehensive AI bill. Passed the Senate, the Chamber is nextBill or draft
- 2025BrazilPL 6237/2025 (National AI System, merged into PL 2338) and PL 5005/2025 (deepfakes)Bill or draft
- Feb 2026BrazilPL 728/2026, algorithmic transparency in public administrationBill or draft
- 15 Jun 2026CanadaProtecting Privacy and Consumer Data Act tabledBill or draft
- May 2024ChileDraft AI legislation, risk-basedBill or draft
- 7 May 2025ColombiaGovernment AI bill PL 442/2025 (Senate) filed. Archived in June 2025 without a first debateBill or draft
- 21 Jul 2026ColombiaPL 025/2026 (House): risk classes, fines, the SIC as fallback enforcer. First-debate rapporteurs named 29 Aug 2026Bill or draft
- Jul 2026MexicoSeveral AI bills under reviewBill or draft
- 5 Jul 2023PeruLaw 31814 promotes the use of AI, with safeguards for rightsIn force or enacted
- 9 Sep 2025PeruSupreme Decree 115-2025-PCM issues the regulation for Law 31814In force or enacted
- 9 Aug 2024USIllinois signs HB 3773, which regulates AI in employmentIn force or enacted
- 19 May 2025USTAKE IT DOWN Act signed. It covers nonconsensual intimate images, including AI fakesIn force or enacted
- 29 Sep 2025USCalifornia signs SB 53, the Transparency in Frontier AI ActIn force or enacted
- 14 May 2026USColorado signs SB 26-189, which replaces its 2024 AI Act. Its duties apply from 1 Jan 2027Upcoming
- 4 Jun 2026USGreat American AI Act released as a House discussion draftBill or draft
- 25 Jun 2026USHouse Science Committee advances 10 AI bills, and an 11th on 21 Jul. None has reached the floorBill or draft
Europe
The EU AI Act entered into force on 1 Aug 2024. Prohibitions and AI literacy duties applied from 2 Feb 2025, and rules for general-purpose AI models from 2 Aug 2025. Article 50 transparency duties apply from 2 Aug 2026, and systems already on the market get until 2 Dec 2026 to mark AI-generated content. The Digital Omnibus moved the Annex III high-risk rules to 2 Dec 2027 and the Annex I rules to 2 Aug 2028. It also adds a ban on AI that generates nonconsensual intimate images or child sexual abuse material. Ireland's Regulation of Artificial Intelligence Act 2026 took effect on 31 Jul 2026 and set up a national AI Office, so the member-state layer is now forming underneath the EU act. The UK has no AI statute and relies on sector law. A private member's AI bill in the Lords never had government backing and fell when the 2024 to 2026 session of Parliament ended.
- 1 Aug 2024EUEU AI Act enters into forceIn force or enacted
- 2 Feb 2025EUAI Act prohibitions and AI literacy duties applyIn force or enacted
- 27 Jul 2026EUDigital Omnibus (Regulation (EU) 2026/1744) enters into force and moves the high-risk datesIn force or enacted
- 2 Dec 2026EUAI Act prohibitions extend to nonconsensual intimate media and child sexual abuse materialUpcoming
- 2 Dec 2027EUAI Act high-risk rules apply to Annex III systems, including credit scoring, life and health insurance pricing and hiringUpcoming
- 31 Jul 2026IrelandRegulation of Artificial Intelligence Act 2026 in force. It sets up the AI Office of IrelandIn force or enacted
- 19 Jun 2025UKData (Use and Access) Act 2025 enacted. Its automated decision-making rules apply from 5 Feb 2026In force or enacted
Asia-Pacific
China has the most binding AI rules, written by AI type: algorithms, deep synthesis, generative AI, content labeling and now human-like interactive services. Japan passed a light-touch AI Promotion Act in May 2025, and it carries no penalties. Taiwan's AI Basic Act took effect on 14 Jan 2026 as a framework for later rules, and South Korea's AI Basic Act on 22 Jan 2026. Vietnam's AI law took effect on 1 Mar 2026 with a grace period. India's November 2025 guidelines are voluntary, and since 20 Feb 2026 its IT Rules have also required platforms to label AI-generated content. Singapore and Australia lead with voluntary frameworks and national plans, and Australia's December 2025 plan dropped the mandatory guardrails it had proposed in favor of existing law. Thailand put a draft AI Act out for public comment in July 2026.
- 1 Mar 2022ChinaAlgorithmic Recommendation Provisions take effectIn force or enacted
- 10 Jan 2023ChinaDeep Synthesis Provisions take effectIn force or enacted
- 15 Aug 2023ChinaGenerative AI Interim Measures take effectIn force or enacted
- 1 Sep 2025ChinaAI content labeling measures take effectIn force or enacted
- 15 Jul 2026ChinaInterim measures for human-like interactive AI services take effectIn force or enacted
- 2024IndiaRegulation of Deepfake Bill introducedBill or draft
- 2025IndiaAI (Ethics and Accountability) Bill introducedBill or draft
- 20 Feb 2026IndiaIT Rules amendment takes effect: platforms must label AI-generated content and attach provenance dataIn force or enacted
- May 2025JapanAI Promotion Act enacted, a light-touch lawIn force or enacted
- Jul 2025New ZealandBiometric Processing Privacy Code 2025In force or enacted
- May 2026New ZealandSocial Security (Modernisation) Amendment enacted, allows automated decisionsIn force or enacted
- 22 Jan 2026South KoreaAI Basic Act takes effectIn force or enacted
- 14 Jan 2026TaiwanAI Basic Act promulgated and in force. The legislature passed it on 23 Dec 2025In force or enacted
- Late 2025VietnamAI law draftedBill or draft
- 1 Mar 2026VietnamLaw on AI takes effect, with a grace period for existing systemsIn force or enacted
- 2 Jul 2026ThailandDraft AI Act published for public hearing: risk tiers, strict liability, administrative finesBill or draft
05 · Instruments
Tools
We sorted each government's instruments into three kinds. Enforceable rules carry legal duties. Suggestive guidance sets the expectations that regulators and customers point to when they ask how an AI system is governed. Enabling policy, mostly funding and national programs, speeds up adoption. Many jurisdictions run all three at once.
Each row below is one jurisdiction. The levels are our own reading.
| Jurisdiction | Enforceable | Suggestive | Enabling | In plain terms |
|---|---|---|---|---|
| Americas | ||||
| Argentina | Argentina, Enforceable: none noted | Argentina, Suggestive: partial | Argentina, Enabling: strong | Bills in Congress, public-sector guide |
| Brazil | Brazil, Enforceable: none noted | Brazil, Suggestive: partial | Brazil, Enabling: strong | Senate passed PL 2338/2023 |
| Canada | Canada, Enforceable: partial | Canada, Suggestive: strong | Canada, Enabling: strong | Federal automated-decision directive |
| Chile | Chile, Enforceable: none noted | Chile, Suggestive: partial | Chile, Enabling: strong | Draft law, national policy |
| Colombia | Colombia, Enforceable: none noted | Colombia, Suggestive: partial | Colombia, Enabling: partial | PL 025/2026 awaits first debate |
| Mexico | Mexico, Enforceable: none noted | Mexico, Suggestive: none noted | Mexico, Enabling: partial | Many bills, national AI lab plan |
| Peru | Peru, Enforceable: partial | Peru, Suggestive: none noted | Peru, Enabling: partial | Law 31814, no penalties of its own, duties phase in from Sep 2026 |
| US | US, Enforceable: partial | US, Suggestive: strong | US, Enabling: strong | State laws, voluntary NIST framework |
| Europe | ||||
| EU | EU, Enforceable: strong | EU, Suggestive: strong | EU, Enabling: strong | AI Act, high-risk rules from Dec 2027 |
| UK | UK, Enforceable: partial | UK, Suggestive: strong | UK, Enabling: strong | Sector law, no government AI bill |
| Ireland | Ireland, Enforceable: strong | Ireland, Suggestive: partial | Ireland, Enabling: strong | AI Act 2026 enforces the EU AI Act nationally |
| Asia-Pacific | ||||
| Australia | Australia, Enforceable: partial | Australia, Suggestive: strong | Australia, Enabling: strong | Existing law, voluntary standard, National AI Plan |
| Bangladesh | Bangladesh, Enforceable: none noted | Bangladesh, Suggestive: none noted | Bangladesh, Enabling: partial | Strategy 2026 to 2030, draft policy |
| China | China, Enforceable: strong | China, Suggestive: partial | China, Enabling: strong | Binding rules by AI type |
| Hong Kong | Hong Kong, Enforceable: partial | Hong Kong, Suggestive: strong | Hong Kong, Enabling: partial | Sector law plus frameworks |
| India | India, Enforceable: partial | India, Suggestive: strong | India, Enabling: strong | IT Rules require AI-content labels, guidelines Nov 2025 |
| Indonesia | Indonesia, Enforceable: none noted | Indonesia, Suggestive: partial | Indonesia, Enabling: partial | Non-binding circular, roadmap |
| Japan | Japan, Enforceable: partial | Japan, Suggestive: strong | Japan, Enabling: strong | Light-touch AI Promotion Act, no penalties |
| New Zealand | New Zealand, Enforceable: partial | New Zealand, Suggestive: strong | New Zealand, Enabling: strong | Privacy Act applies, light touch |
| Singapore | Singapore, Enforceable: none noted | Singapore, Suggestive: strong | Singapore, Enabling: strong | Voluntary frameworks, incl. agentic AI |
| South Korea | South Korea, Enforceable: strong | South Korea, Suggestive: strong | South Korea, Enabling: strong | AI Basic Act |
| Taiwan | Taiwan, Enforceable: partial | Taiwan, Suggestive: partial | Taiwan, Enabling: strong | AI Basic Act, a framework law with no penalties |
| Thailand | Thailand, Enforceable: none noted | Thailand, Suggestive: partial | Thailand, Enabling: partial | Draft AI Act out for comment, ethics guideline |
| Vietnam | Vietnam, Enforceable: strong | Vietnam, Suggestive: none noted | Vietnam, Enabling: strong | AI law in effect since March 2026 |
06 · Overlap
Common ground
We counted how many jurisdictions have each requirement in our review. These counts are a floor, because the review is a summary.
- Data protection law in force23 of 24
Argentina, Australia, Bangladesh, Brazil, Canada, Chile, China, Colombia, EU, Hong Kong, India, Indonesia, Ireland, Japan, Mexico, New Zealand, Peru, Singapore, South Korea, Taiwan, Thailand, UK, Vietnam
- Transparency and disclosure17 of 24
Argentina, Australia, Brazil, Canada, Chile, China, EU, India, Japan, New Zealand, Peru, Singapore, South Korea, Taiwan, Thailand, UK, Vietnam
- Risk tiers for high-risk AI13 of 24
Argentina, Australia, Brazil, Canada, Chile, Colombia, EU, New Zealand, South Korea, Taiwan, Thailand, US, Vietnam
- Safety testing and security10 of 24
Australia, Brazil, Canada, China, India, Japan, Singapore, South Korea, UK, US
- AI guidance from data regulators10 of 24
Argentina, Australia, Brazil, Canada, Hong Kong, Ireland, New Zealand, Singapore, South Korea, UK
- Accountability and liability8 of 24
Australia, Brazil, Colombia, EU, India, Japan, Peru, Singapore
- Public-sector use rules7 of 24
Argentina, Australia, Brazil, Canada, New Zealand, UK, US
- Deepfake and synthetic content7 of 24
Brazil, China, EU, India, Mexico, South Korea, US
- A named agentic AI framework1 of 24
Singapore
07 · Sources
Method and sources
We reviewed each jurisdiction's AI laws, bills, guidance and strategies from official texts where they were available, and from legal analysis where they were not. Every dated item was checked against its source on 17 Oct 2026. An obligation with a future start date appears on the timeline but is left out of the counts. The main sources:
- EU: AI Act, Regulation (EU) 2024/1689; Digital Omnibus, Regulation (EU) 2026/1744; Code of Practice on AI-generated content
- Ireland and UK: Regulation of Artificial Intelligence Bill 2026; Data (Use and Access) Act 2025 guidance; Artificial Intelligence (Regulation) Bill [HL]
- US: Executive Order 14365; Great American AI Act discussion draft; House Science Committee markup; California SB 53; Colorado SB 26-189
- Canada and Latin America: Canada's privacy bill; Brazil PL 2338/2023; Colombia's AI bills; Peru Supreme Decree 115-2025-PCM
- Asia-Pacific: India IT Rules amendment; Japan AI Promotion Act; Taiwan AI Basic Act; Thailand draft AI Act; Australia National AI Plan
- Cross-checks: IAPP Global AI Law and Policy Tracker; OECD.AI policy dashboards
Not legal advice.