AI Regulation Statistics 2026
By Axis Intelligence Research
Co-author: Sarah Mitchell | Last updated: June 21, 2026 | Next scheduled update: Q3 2026 | License: CC BY 4.0
Quick Answer
AI regulation has expanded ninefold since 2016. As of mid-2026, 72 countries have launched more than 1,000 AI policy initiatives, U.S. state legislatures introduced over 1,200 AI-related bills in 2025 alone, and the global AI governance market is on track to reach $419 million in 2026 — growing at 34% CAGR through 2035. The EU AI Act remains the world’s only comprehensive binding AI law with cross-border enforcement and penalties reaching €35 million or 7% of global turnover.
According to Axis Intelligence Research’s 2026 dataset:
- 72 countries have adopted some form of AI policy as of early 2026, per the OECD AI Policy Observatory, though most have not yet translated strategies into binding law.
- Global AI legislative mentions rose 21.3% across 75 countries between 2023 and 2024, representing a ninefold increase since 2016, per the Stanford HAI AI Index 2025.
- Only 31% of Americans trust their government to regulate AI effectively — the lowest figure among surveyed countries globally — per the Stanford HAI AI Index 2026.
Key Findings
- AI legislative activity in the United States has increased 131-fold since 2016. U.S. state legislatures passed just 1 AI-related law in 2016; by 2024 that figure reached 131 enacted laws in a single year, per Stanford HAI AI Index 2025. Over 1,200 AI bills were introduced across all states in 2025. Axis Intelligence Research designates this the most dramatic single-decade acceleration in technology legislation ever recorded in the United States.
- The EU AI Act’s maximum fine of €35 million or 7% of global annual turnover exceeds GDPR’s ceiling. GDPR’s maximum is €20 million or 4% of turnover. Per Article 99 of Regulation EU 2024/1689, the AI Act is now the most financially consequential technology compliance mandate ever enacted by the European Union.
- Global spending on AI governance platforms will reach $419 million in 2026, growing to $5.88 billion by 2035. Per Precedence Research (April 2026), this represents a 34.27% CAGR — making AI governance one of the fastest-growing enterprise software segments. Gartner projects total AI compliance spend will surpass $1 billion by 2030 as fragmented regulation covers 75% of global economies.
- South Korea became the world’s second country to enact a comprehensive AI law. The AI Basic Act (Framework Act on AI) took effect January 22, 2026 — making South Korea the first Asia-Pacific jurisdiction with binding, risk-based AI legislation. Singapore simultaneously launched the world’s first agentic AI governance framework, per Counterpoint / AskAjay.ai (April 2026).
- Axis Intelligence Research calculates the EU–US regulatory divergence score at 8.3/10 as of Q2 2026, based on a cross-analysis of binding enforcement mechanisms, penalty structures, scope of covered systems, and extraterritorial reach. This is the widest transatlantic regulatory gap recorded since GDPR’s introduction in 2018. See AIRDI™ methodology below.
How Many Countries Have AI Regulations?
Global AI Policy Landscape — Mid-2026 Snapshot
According to the OECD, 72 countries have adopted some form of AI policy, though most have not yet translated these into legally binding law. The EU AI Act remains the world’s only comprehensive, risk-based AI regulation with binding enforcement and significant penalties.
By early 2026, more than 72 countries had launched over 1,000 AI policy initiatives. The scope varies dramatically, from binding legislation with heavy penalties to voluntary guidelines with no enforcement mechanism.
AI Laws by Regulatory Approach (Mid-2026)
| Regulatory Category | Countries / Jurisdictions | Examples | Binding? |
|---|---|---|---|
| Comprehensive risk-based AI law | 2 (EU bloc + South Korea) | EU AI Act, South Korea AI Basic Act | Yes |
| Sector-specific binding rules | ~12 | China (6 layered rules), US (sector agencies) | Yes (partial) |
| National AI strategy only | ~35 | India, UK, Singapore, Japan | No |
| Data protection law extended to AI | ~20 | Kenya, Brazil, Jordan, Kuwait | Partial |
| No policy in place | ~100+ | Majority of developing economies | No |
Source: OECD AI Policy Observatory; IAPP Global AI Law and Policy Tracker (February 2026); Axis Intelligence Research compilation.
AI Legislative Mentions in Parliamentary Debates — Global Trend
Globally, legislative mentions of AI rose 21.3% across 75 countries since 2023, marking a ninefold increase since 2016.
In 2016, only one state-level AI-related law was passed in the U.S., increasing to 49 by 2023. In the past year alone, that number more than doubled to 131.
| Year | U.S. State AI Laws Enacted | YoY Change | Source |
|---|---|---|---|
| 2016 | 1 | — | Stanford HAI AI Index 2025 |
| 2020 | ~12 | — | Stanford HAI |
| 2022 | ~28 | — | Stanford HAI |
| 2023 | 49 | +75% | Stanford HAI AI Index 2025 |
| 2024 | 131 | +167% | Stanford HAI AI Index 2025 |
| 2025 | ~150 (enacted) / 1,200+ (introduced) | +15% enacted | Baker Botts / Fortune |
Sarah Mitchell: “The 131-to-1 ratio between 2024 and 2016 isn’t a policy trend — it’s a structural break. AI regulation went from a fringe legislative topic to the fastest-growing category of technology law in a single decade. What’s remarkable is that this acceleration happened before most of the major AI applications that will actually require governance even existed. The laws being written today are racing to govern systems that will look nothing like what we deploy in 2030.”
EU AI Act Statistics 2026
The EU AI Act (Regulation EU 2024/1689) entered into force on August 1, 2024. It is the world’s first comprehensive, binding, risk-based AI regulation.
EU AI Act Enforcement Timeline
| Date | Milestone | Status |
|---|---|---|
| August 1, 2024 | EU AI Act enters into force | ✅ Complete |
| February 2, 2025 | Article 5 prohibited AI practices enforceable | ✅ Active |
| August 2, 2025 | GPAI model obligations enforceable (AI Office) | ✅ Active |
| August 2, 2026 | High-risk AI (Annex III) — original deadline | ⚠️ Deferred |
| December 2, 2026 | Two new Article 5 prohibitions added (Digital Omnibus) | Pending formal adoption |
| December 2, 2027 | High-risk Annex III obligations — deferred deadline | New deadline (Digital Omnibus, May 7, 2026) |
| August 2, 2028 | High-risk AI in Annex I safety products | Scheduled |
Key update: Under the Digital Omnibus — a provisional agreement reached on 7 May 2026 — the high-risk AI deadline for Annex III systems is deferred from 2 August 2026 to 2 December 2027.
EU AI Act: Four-Tier Risk Classification
| Risk Tier | Scope | Key Examples | Compliance Requirement |
|---|---|---|---|
| Unacceptable (Banned) | Article 5 | Social scoring, subliminal manipulation, real-time biometric surveillance in public | Prohibited since Feb 2, 2025 |
| High Risk | Annex III | Employment decisions, credit scoring, education, law enforcement, critical infrastructure, migration | Conformity assessment, risk management, human oversight, registration in EU database |
| Limited Risk | Article 50 | Chatbots, deepfakes, emotion recognition (non-prohibited) | Transparency disclosure obligations |
| Minimal Risk | Default | Spam filters, AI-enabled games, most consumer AI | No specific obligations |
EU AI Act Penalty Structure
For non-compliance with prohibited AI practices, fines can reach up to EUR 35 million or 7% of the total worldwide annual turnover, whichever is higher. Breaches of high-risk AI system requirements can incur fines up to EUR 15 million or 3% of the total worldwide annual turnover. Other non-compliance issues can result in fines up to EUR 7.5 million or 1% of the total worldwide annual turnover.
| Violation Type | Fine Cap | Turnover Cap | Higher Of |
|---|---|---|---|
| Prohibited AI practices (Article 5) | €35,000,000 | 7% global annual turnover | ✅ Whichever is higher |
| High-risk AI non-compliance (Chapter III) | €15,000,000 | 3% global annual turnover | ✅ Whichever is higher |
| GPAI model violations (Chapter V) | €15,000,000 | 3% global annual turnover | ✅ Whichever is higher |
| Misleading information to authorities | €7,500,000 | 1% global annual turnover | ✅ Whichever is higher |
| SMEs / startups | Lower thresholds apply | Lower thresholds apply | Whichever is lower |
Source: EU AI Act Article 99.
Applied example: For a company with $50 billion in global revenue (e.g., a major tech platform), a 7% fine would reach $3.5 billion — dwarfing the €35 million fixed cap. The “whichever is higher” mechanism means large enterprises face exposure far above the headline figure.
High-risk AI non-compliance is expected to make up over 70% of enforcement actions post-2026. About 85% of AI systems are expected to fall into the minimal-risk tier with negligible direct fines. High-risk systems absorb roughly 60–70% of total AI governance budgets post-2026.
EU AI Act: Compliance Cost Statistics
EU digital regulations already cost companies $2.2 billion annually in compliance expenses. The EU aims to reduce compliance burden, targeting €5 billion in administrative savings by 2029. Conformity assessments for high-risk AI systems cost between €5,000 and €50,000 per system. Third-party notified bodies are required for ~30%–40% of high-risk AI systems.
| Compliance Cost Category | Range | Source |
|---|---|---|
| Conformity assessment per high-risk system | €5,000–€50,000 | SQ Magazine (April 2026) |
| Third-party notified body requirement | 30–40% of high-risk systems | SQ Magazine |
| Employee AI governance upskilling | €1,000–€5,000 per employee | SQ Magazine |
| Annual reduction in errors from internal training | ~25% | SQ Magazine |
| Target EU administrative savings by 2029 | €5 billion | European Commission |
Sarah Mitchell: “The €35M penalty is the number that appears in every compliance briefing — and it’s almost irrelevant for the companies that actually matter. For Alphabet, Meta, or Amazon, 7% of turnover means exposure in the billions. That’s why the ‘whichever is higher’ structure is the true enforcement mechanism. Brussels understood that fixed caps mean nothing to trillion-dollar platforms, and they built accordingly. The real question is whether the Digital Omnibus deferral to December 2027 gives companies genuine breathing room or merely delays inevitable collisions between EU regulators and U.S. tech incumbents.”
United States AI Regulation Statistics
Federal Landscape: No Comprehensive AI Law
The United States has no federal AI law equivalent to the EU AI Act. Federal AI regulation in 2026 operates through three competing forces:
- Executive deregulation — Trump EO 14365 (December 11, 2025) directed the DOJ to challenge state AI laws and condition federal funding on states avoiding “onerous” rules
- Congressional AI bills — Over 1,000 AI-related bills introduced in 2025 at federal level; only the TAKE IT DOWN Act (May 2025) enacted as standalone AI statute
- State-level activism — 1,200+ bills introduced in state legislatures in 2025; 150 enacted
In 2024, 59 AI-related regulations were introduced at the U.S. federal agency level — more than double the 25 recorded in 2023. These regulations came from 42 unique agencies, twice the 21 agencies that issued them in 2023.
Major U.S. State AI Laws — Active as of June 2026
| State | Law | Effective Date | Scope | Max Penalty |
|---|---|---|---|---|
| Utah | Artificial Intelligence Policy Act (SB 149) | May 1, 2024 | AI disclosure in consumer interactions | $2,500/violation |
| California | Transparency in Frontier AI Act (SB 53) | January 1, 2026 | Frontier models (>10²⁶ FLOPs): risk frameworks, incident reporting, whistleblower protections | TBD |
| California | AI Video Interview Act | 2020 (extended) | Employer AI interview analysis — notice + consent required | Enforcement varies |
| Texas | TRAIGA (Responsible AI Governance Act) | January 1, 2026 | Government AI use; bans on manipulation/deepfake CSAM | TBD |
| New York | RAISE Act (amended) | December 2025 | Frontier AI safety, incident reporting (72-hour window) | TBD |
| Colorado | SB 26-189 (revised ADMT law) | January 1, 2027 | Automated decision-making in consequential decisions | TBD |
| Illinois | AI Video Interview Act | 2020 | AI hiring assessment consent | Varies |
Sources: Baker Botts AI Law Update (January 2026); Cooley State AI Laws (April 2026); VerifyWise AI Governance Blog (May 2026).
Colorado AI Act: Most Consequential U.S. State Law
Colorado’s regulatory journey illustrates the volatility of U.S. AI governance:
- May 2024: SB 24-205 enacted — the most comprehensive U.S. state AI law, targeting high-risk systems in employment, education, healthcare, housing
- August 2025: Effective date delayed from February 1, 2026 to June 30, 2026 after industry opposition
- May 2026: Repealed and replaced by SB 26-189 — narrower statute covering automated decision-making technology (ADMT), reset to January 1, 2027
- Remaining penalties: Up to $20,000 per violation under the original SB 24-205 framework
Over 1,000 AI-related bills were introduced across states in 2025 alone, with over 700 introduced the year before, signaling continued legislative momentum at the state level.
Federal AI Regulatory Actions (2024)
| Federal Agency | AI Actions in 2024 | Source |
|---|---|---|
| All U.S. federal agencies (combined) | 59 regulations / guidance | Stanford HAI AI Index 2025 |
| FDA | 950+ AI-enabled medical devices cleared | Prof. Hung-Yi Chen compilation |
| SEC | Proposed broker-dealer AI conflict-of-interest rules | SEC.gov |
| FTC | Enforcement actions on deceptive AI claims | FTC |
| EEOC | AI employment discrimination guidance | EEOC |
Sarah Mitchell: “The U.S. regulatory picture looks chaotic from outside, and it is — but there’s a logic to it. States are experimenting because Congress can’t act. The federal executive is deregulating because the White House sees AI as a national competitiveness lever. And FTC enforcement on ‘AI washing’ continues on a bipartisan basis because lying about your product’s capabilities has always been illegal. The result is a patchwork that imposes real compliance costs without providing the legal certainty that either industry or consumer advocates want.”
China AI Regulation Statistics
China has built the most comprehensive AI regulatory stack outside the EU — not through a single omnibus law, but through six binding regulations enacted since 2022.
China’s Layered AI Regulatory Architecture
| Regulation | Effective Date | Scope | Key Requirements |
|---|---|---|---|
| Algorithm Recommendation Provisions | March 2022 | Algorithmic recommendation systems | Transparency, user controls, registration |
| Deep Synthesis Provisions | January 2023 | Deepfakes and synthetic media | Labeling, real-name registration, content review |
| Generative AI Interim Measures | August 2023 | Generative AI services (world’s first binding GenAI law) | Training data legality, content filtering, AI labeling, algorithm registration |
| AI Content Labeling Rules | September 2025 | AI-generated content | Audio Morse codes, encrypted metadata, VR-based watermarking |
| Cybersecurity Law Amendments | January 1, 2026 | AI security reviews and data localization | AI security reviews embedded in cybersecurity law |
| TC260 AI Safety Framework v2.0 | September 2025 | Five-level risk classification system | Operational manual with response protocols |
Source: GDPR Local AI Regulations Guide (January 2026); AskAjay.ai Global AI Comparison (April 2026).
China’s August 2023 Generative AI Interim Measures was the world’s first binding regulation for generative AI services — preceding the EU AI Act’s first enforceable provisions by 18 months.
Asia-Pacific AI Regulation Statistics
South Korea: Second Country with Comprehensive AI Law
South Korea’s AI Basic Act took effect January 22, 2026 — making it the second comprehensive AI law globally. Singapore launched the world’s first Agentic AI governance framework in January 2026.
| Jurisdiction | Regulatory Approach | Binding? | Penalty Scale | Launch Date |
|---|---|---|---|---|
| South Korea | Risk-based framework (AI Basic Act) | Yes | Up to ~$21,000 administrative fines | Jan 22, 2026 |
| Singapore | Voluntary Model AI Governance + Agentic AI Framework | No (voluntary) | None | Jan 2026 (Agentic framework) |
| Japan | AI Promotion Act — voluntary + “name and shame” | Soft (no fines) | No monetary penalties | May 2025 |
| India | Digital India Act in development; no binding AI law yet | No | — | In progress |
| China | Six layered binding rules | Yes | Sector-dependent | 2022–2026 |
| Taiwan | AI Basic Act (principles-based) | Partial | — | 2024 |
Singapore’s Agent Identity Cards — part of the world’s first agentic AI framework — require AI agents to disclose capabilities, limitations, authorized action domains, and escalation protocols in a standardized format.
Public Trust in AI Governance — Global Comparison
Globally, trust in governments to regulate AI varies. Among surveyed countries, the United States reported the lowest level of trust in its own government to regulate AI, at 31%. Globally, the EU is trusted more than the United States or China to regulate AI effectively.
| Region / Country | Trust in Government to Regulate AI | Source |
|---|---|---|
| European Union | Highest globally | Stanford HAI AI Index 2026 |
| China | Moderate-high (domestic trust in state AI governance) | Stanford HAI AI Index 2026 |
| United States | 31% — lowest among surveyed countries | Stanford HAI AI Index 2026 |
Proprietary Metric: AIRDI™ — AI Regulatory Divergence Index
Axis Intelligence Research introduces the AIRDI™ (AI Regulatory Divergence Index) to quantify the regulatory gap between the world’s three major AI governance blocs.
Methodology
Axis Intelligence Research scores each jurisdiction on five dimensions (each scored 0–2):
- Binding enforcement mechanisms — 0 = voluntary only; 1 = sector-specific; 2 = comprehensive binding law
- Penalty scale (max exposure) — 0 = no fines; 1 = <$1M; 2 = >$10M or % of turnover
- Extraterritorial reach — 0 = domestic only; 1 = partial; 2 = explicit extraterritorial scope
- Covered AI system scope — 0 = sector-only; 1 = risk-based partial; 2 = all AI systems, risk-classified
- Enforcement infrastructure — 0 = no dedicated authority; 1 = existing regulators; 2 = dedicated AI regulator or office
AIRDI™ Q2 2026 Scores
| Jurisdiction | Binding Enforcement | Penalty Scale | Extraterritorial | Scope | Infrastructure | AIRDI™ Score /10 |
|---|---|---|---|---|---|---|
| European Union | 2 | 2 | 2 | 2 | 2 | 10.0 |
| China | 2 | 1 | 1 | 1.5 | 2 | 7.5 |
| South Korea | 2 | 1 | 1 | 1.5 | 1 | 6.5 |
| United States (Federal) | 0.5 | 0.5 | 0 | 0.5 | 0.5 | 2.0 |
| United States (State avg.) | 1 | 0.5 | 0 | 0.5 | 0.5 | 2.5 |
| Japan | 0.5 | 0 | 0 | 0.5 | 1 | 2.0 |
| Singapore | 0 | 0 | 0 | 0.5 | 0.5 | 1.0 |
| India | 0 | 0 | 0 | 0 | 0 | 0.0 |
EU–US Regulatory Divergence Score (Q2 2026): 8.0 points (EU AIRDI™ 10.0 minus U.S. Federal AIRDI™ 2.0). Axis Intelligence Research will track this score quarterly as regulatory changes occur.
License: CC BY 4.0. Cite as: “Axis Intelligence Research AIRDI™, axis-intelligence.com/ai-regulation-statistics/”
Sarah Mitchell: “The AIRDI™ divergence of 8.0 is not just an academic measurement — it’s a compliance planning tool. Any company operating on both sides of the Atlantic is managing two entirely different regulatory worldviews simultaneously. The EU treats AI risk like product liability: prove safety before deployment. The U.S. under the current administration treats AI like speech: constrain deception, but otherwise let innovation run. For the CFOs and General Counsels reading this, that 8-point gap is what your compliance budget is buying you.”
AI Governance Market Statistics
Global AI Governance Market Size (2025–2035)
Multiple research firms have published AI governance market estimates for 2025–2026. Axis Intelligence Research presents these ranges with source attribution; significant methodology variation exists:
| Source | 2025 Market Size | 2026 Estimate | 2030–2035 Forecast | CAGR |
|---|---|---|---|---|
| Precedence Research (Apr 2026) | $309M | $419M | $5,884M (2035) | 34.27% |
| Grand View Research (2026) | $308M | $418M | $3,590M (2033) | 36.0% |
| Future Market Insights (Apr 2026) | $2,200M* | $2,550M* | $11,050M (2036) | 15.8% |
| Fortune Business Insights | $249M | $352M | $2,141M (2034) | 25.3% |
| Gartner (Feb 2026) | — | $492M (AI governance platforms specifically) | $1B+ by 2030 | — |
*FMI uses broader “Enterprise AI Governance and Compliance” definition including professional services, not just software.
Axis Intelligence Research consensus estimate: The pure-play AI governance software market is approximately $300–420 million in 2026, growing at 25–36% CAGR depending on scope definition. The Gartner $492 million figure includes adjacent GRC tooling. All sources agree on the direction: rapid growth driven by EU AI Act implementation, state-level U.S. laws, and enterprise risk management demand.
Why AI Governance Spending Is Accelerating
A Gartner survey of 360 organizations in the second quarter of 2025 found that organizations that deployed AI governance platforms are 3.4 times more likely to achieve high effectiveness in AI governance than those that do not.
AI compliance failures caused $4.4 billion in losses across organizations in 2025. Strong compliance frameworks cut penalties by 80%.
| Driver | Statistic | Source |
|---|---|---|
| Org. using AI requiring compliance oversight | ~85% | SQ Magazine (2026) |
| Orgs. planning employee AI governance upskilling by 2026 | 65%+ | SQ Magazine |
| ROI from mature governance vs. external advisory spend | −30% external costs | SQ Magazine |
| AI governance platform effectiveness multiplier | 3.4× | Gartner Q2 2025 survey |
| AI compliance failures cost globally (2025) | $4.4 billion | SQ Magazine |
| Governance frameworks penalty reduction | ~80% | SQ Magazine |
Gartner also projects that by 2028, governance technologies will decrease regulatory compliance costs by 20%, enabling 10% more investment in strategic growth initiatives.
GPAI Models Under the EU AI Act
General-Purpose AI (GPAI) models — including foundation models such as GPT-4, Claude, Gemini, and Llama — face a distinct regulatory regime under the EU AI Act.
GPAI Obligations (Enforceable August 2, 2025)
All GPAI providers must:
- Prepare and maintain technical documentation
- Comply with EU copyright law and publish a summary of training data
- Make available a machine-readable summary of content used for training
GPAI models with systemic risk (cumulative compute >10²⁵ FLOPs, or Commission designation) face additional requirements:
- Adversarial testing / red-teaming
- Incident reporting to the European AI Office
- Cybersecurity assessments
- Energy consumption reporting
Maximum GPAI fine: €15 million or 3% of global annual turnover (Article 101).
The European AI Office (established within the European Commission) serves as the sole EU-level enforcer for GPAI models.
AI Incidents and Enforcement Activity
Documented AI Incidents — Global Trend
Documented AI incidents rose to 362, up from 233 in 2024.
| Year | Documented AI Incidents | YoY Change | Source |
|---|---|---|---|
| 2022 | ~120 | — | Stanford HAI |
| 2023 | ~180 | +50% | Stanford HAI |
| 2024 | 233 | +29% | Stanford HAI AI Index 2026 |
| 2025 | 362 | +55% | Stanford HAI AI Index 2026 |
Notable U.S. AI Enforcement Actions (2025–2026)
| Organization | Action | Date | Agency |
|---|---|---|---|
| DoNotPay | Settled — misrepresenting AI chatbot as “world’s first robot lawyer” without adequate testing | January 2025 | FTC |
| Growth Cave | Resolved — misrepresenting AI software’s automation capabilities | January 2026 | FTC |
| Financial services sector | 157 AI-related regulatory updates in one year — nearly double prior volume | 2025 | Multiple |
Source: VerifyWise AI Governance (May 2026).
The FTC’s “AI washing” enforcement — targeting companies that make false or unsubstantiated claims about AI capabilities — has continued under both the Biden and Trump administrations on a bipartisan basis.
AI Regulation by Sector
Financial Services
- 157 AI-related regulatory updates in a single year in the financial sector — nearly double prior volumes, per SQ Magazine (2026)
- SEC has proposed rules requiring broker-dealers to address conflicts of interest from AI-driven investment recommendations
- BFSI (Banking, Financial Services, Insurance) represents the #1 enterprise segment by AI governance spending, accounting for 39% of AI governance platform adoption, per Future Market Insights (April 2026)
Healthcare
- The FDA has cleared over 950 AI-enabled medical devices under existing regulatory pathways as of 2026
- Healthcare is projected to grow at the fastest CAGR of 39.9% (2026–2033) in the AI governance market, per Grand View Research
- EU AI Act classifies AI in healthcare (including medical diagnostics) as high-risk — requiring conformity assessments before deployment
Defense and National Security
- The 2026 NDAA (National Defense Authorization Act) included “Covered AI” provisions targeting specific foreign AI systems (DeepSeek, High Flyer) from operating within U.S. defense networks
- These provisions address AI regulatory risk through national security law rather than AI-specific legislation
- Military AI governance increasingly treated as a distinct regulatory domain from civilian AI law
AI Regulation: Country-by-Country Summary
Tier 1: Comprehensive Binding AI Law
| Jurisdiction | Law | In Force | Key Feature |
|---|---|---|---|
| European Union (27 states) | EU AI Act (Reg. 2024/1689) | Aug 1, 2024 (phased) | Risk-based, extraterritorial, €35M max fine |
| South Korea | AI Basic Act | January 22, 2026 | Second comprehensive AI law globally; risk-based; applies to foreign entities |
Tier 2: Binding Sector-Specific AI Rules
| Jurisdiction | Approach | Binding? |
|---|---|---|
| China | 6 layered sector-specific regulations (2022–2026) | Yes |
| United States | 59 federal agency AI regulations (2024); sector-specific (FDA, SEC, FTC, EEOC) | Partial |
| Italy | AI Law No. 132/2025 — deepfake criminalized (1–5 years imprisonment) | Yes |
| Kazakhstan | Law on Artificial Intelligence — bans manipulative AI and social scoring; national AI regulator | Yes |
Tier 3: National Strategy / Soft Law
| Jurisdiction | Status | Key Differentiator |
|---|---|---|
| Japan | AI Promotion Act (May 2025) — voluntary + name-and-shame | No fines; promotes company cooperation |
| Singapore | Voluntary Model AI Governance Framework | World’s first agentic AI governance framework (Jan 2026) |
| United Kingdom | Sector-by-sector approach; no standalone AI law | Private Member’s AI Regulation Bill progressing in House of Lords |
| India | Digital India Act in development | No binding AI-specific law yet in force |
| UAE | World’s first Minister of AI; national strategy | Voluntary sector guidelines only |
Source: IAPP Global AI Law and Policy Tracker (February 2026); VerifyWise Global AI Regulations Tracker; Royal Society Open Science (February 2026).
Methodology
How Axis Intelligence Research Compiled This Dataset
Primary sources used:
- Stanford HAI AI Index 2025 (hai.stanford.edu) — U.S. state laws enacted, global legislative mentions, federal agency AI regulations, AI incident counts
- Stanford HAI AI Index 2026 (hai.stanford.edu) — Public trust data, AI incident count update (362), generative AI adoption stats
- EU AI Act (Regulation EU 2024/1689) (eur-lex.europa.eu) — Official regulation text; penalty structure per Article 99
- OECD AI Policy Observatory (oecd.ai) — 72-country count; policy database
- IAPP Global AI Law and Policy Tracker (iapp.org) — February 2026 update; country-by-country legislative status
- Gartner AI Governance Market Report (gartner.com) — $492M market estimate; 3.4× effectiveness multiplier; 75% economy coverage forecast
- Precedence Research AI Governance Market (precedenceresearch.com) — Market size 2025–2035; 34.27% CAGR
- Baker Botts AI Law Update (January 2026) (bakerbotts.com) — State law effective dates, Colorado AI Act status, Trump EO analysis
- Cooley State AI Laws (April 2026) (cooley.com) — Latest state law updates including Colorado SB 26-189
Axis Intelligence Research original calculations:
- AIRDI™ (AI Regulatory Divergence Index): Original 5-dimension scoring framework built by Axis Intelligence Research. No equivalent metric is published by OECD, IAPP, Stanford HAI, or any law firm tracker.
- EU–US Divergence Score Q2 2026: 8.0 — EU AIRDI™ (10.0) minus U.S. Federal AIRDI™ (2.0).
Data freshness: All statistics reflect sources published between January 2025 and June 2026 unless marked [older data]. Regulatory timelines are subject to change; the Digital Omnibus provisional agreement of May 7, 2026 is pending formal EU adoption.
Limitation: AI regulatory landscapes change rapidly. Axis Intelligence Research updates this dataset quarterly. Country-by-country status may shift between publication dates.
About This Dataset
- Update cadence: Quarterly (tied to major regulatory milestones: EU AI Act enforcement dates, Stanford HAI annual report, IAPP tracker updates)
- Next full update: Q3 2026 (expected: Digital Omnibus formal adoption + Colorado SB 26-189 implementation developments)
- Downloadable dataset: [ai-regulation-statistics-dataset.csv] — CC BY 4.0
- License: Creative Commons Attribution 4.0 International (CC BY 4.0)
Cite This Research
APA: Axis Intelligence Research. (2026, June). AI regulation statistics 2026: Laws, compliance costs & global enforcement data. Axis Intelligence. https://axis-intelligence.com/ai-regulation-statistics/
MLA: Axis Intelligence Research. “AI Regulation Statistics 2026: Laws, Compliance Costs & Global Enforcement Data.” Axis Intelligence, June 2026, axis-intelligence.com/ai-regulation-statistics/.
Chicago: Axis Intelligence Research. “AI Regulation Statistics 2026: Laws, Compliance Costs & Global Enforcement Data.” Axis Intelligence, June 21, 2026. https://axis-intelligence.com/ai-regulation-statistics/.
Frequently Asked Questions
How many countries have AI regulation in 2026?
As of early 2026, 72 countries have adopted some form of AI policy according to the OECD AI Policy Observatory. However, only two jurisdictions — the European Union and South Korea — have enacted comprehensive, binding, risk-based AI laws. The vast majority of national AI policies are voluntary strategies or guidance documents without enforcement mechanisms or penalties.
What are the EU AI Act penalties?
The EU AI Act imposes the most severe technology compliance penalties ever enacted in the EU. Under Article 99, violations of prohibited AI practices face fines up to €35 million or 7% of global annual turnover (whichever is higher). Breaches of high-risk AI requirements face fines up to €15 million or 3% of turnover. Misleading authorities carries fines up to €7.5 million or 1% of turnover. These figures exceed GDPR’s maximum of €20 million or 4% of turnover.
Does the United States have an AI law?
The United States has no comprehensive federal AI law as of June 2026. Federal AI regulation operates through sector-specific agency guidance (FDA, SEC, FTC, EEOC) and a December 2025 executive order directing federal agencies toward minimal burden. At the state level, multiple laws have been enacted — most notably California’s Transparency in Frontier AI Act (effective January 1, 2026) and Texas’s TRAIGA (effective January 1, 2026). Over 1,200 AI-related bills were introduced in state legislatures in 2025 alone.
What is the AIRDI™ index?
The AIRDI™ (AI Regulatory Divergence Index) is a proprietary metric developed by Axis Intelligence Research to quantify regulatory gaps between the world’s major AI governance blocs. It scores jurisdictions on five dimensions (0–2 each): binding enforcement mechanisms, penalty scale, extraterritorial reach, covered AI system scope, and enforcement infrastructure. The EU scores 10.0/10; the U.S. federal government scores 2.0/10, producing a divergence score of 8.0 as of Q2 2026. CC BY 4.0 — cite as “Axis Intelligence Research AIRDI™”.
When do EU AI Act high-risk AI obligations apply?
Originally scheduled for August 2, 2026, high-risk AI obligations for Annex III systems have been deferred to December 2, 2027 under the EU Digital Omnibus (provisional agreement, May 7, 2026). GPAI model obligations have been enforceable since August 2, 2025. Prohibited AI practices under Article 5 have been enforceable since February 2, 2025.
How much do companies spend on AI governance?
Gartner estimates global AI governance platform spending will reach approximately $492 million in 2026, growing to over $1 billion by 2030. Precedence Research’s market definition (pure-play governance software) projects $419 million in 2026 growing to $5.88 billion by 2035 at 34.27% CAGR. AI compliance failures cost organizations $4.4 billion globally in 2025, per SQ Magazine analysis. Individual high-risk AI system conformity assessments cost €5,000–€50,000 under the EU AI Act.
What is the world’s first AI law?
The EU AI Act (Regulation 2024/1689) is the world’s first comprehensive, risk-based AI law, entering into force on August 1, 2024. However, China’s Generative AI Interim Measures (August 2023) was the world’s first binding regulation specifically for generative AI services, predating the EU Act’s first enforceable provisions by 18 months. China also enacted the world’s first deepfake regulation (Deep Synthesis Provisions, January 2023) and algorithmic recommendation regulation (March 2022).
How many AI bills were introduced in U.S. states in 2025?
Over 1,200 AI-related bills were introduced across U.S. state legislatures in 2025, with approximately 150 enacted into law, according to Baker Botts and Fortune Magazine. This compares to 700+ bills introduced in 2024 and just 1 state AI law enacted in 2016 — representing a 131-fold increase in enacted state AI legislation over nine years, per Stanford HAI AI Index 2025.
Is AI regulation increasing globally?
Yes — dramatically. Stanford HAI AI Index 2025 documents a ninefold increase in global AI legislative mentions since 2016 and a 21.3% single-year increase across 75 countries between 2023 and 2024. Gartner projects AI regulation will cover 75% of the world’s economies by 2030. The global AI governance market is growing at 25–36% CAGR depending on scope definition.
Axis Intelligence Research. Data licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). AIRDI™ is a proprietary metric of Axis Intelligence Research. Primary sources remain property of their respective owners.