{"content_id":"x6tfpqnhak","slug":"ai-regulation-fragmentation-un-eu-us-states-july-2026","locale":"en","schema_type":"Report","category":"report","category_name":"Report","title":"Fragmentation of AI Regulation by Country: A Map of UN, EU, and U.S. State Regulations as of July 2026","summary":"In July 2026, AI regulation entered a phase of fragmentation, with discussions on international governance at the UN level, adjustments to the schedule and procedures of the EU’s AI Act, and the spread of state-level legislation in the U.S. all proceeding simultaneously. Global AI companies must manage not a single set of rules, but rather region-specific implementation dates, risk classifications, transparency obligations, protections for minors, and restrictions on high-risk uses—just as they manage data.","key_points":["The UN/ITU’s AI for Good Global Commission can be understood as an international governance coordination mechanism that connects corporate CEOs and national policymakers, rather than as a legal framework in and of itself.","While maintaining a risk-based regulatory framework, the EU AI Act is influencing companies’ compliance roadmaps through discussions on streamlining and schedule adjustments set for late June 2026.","In the United States, with no comprehensive federal law in place, state-by-state regulations are spreading in states such as Illinois, Colorado, and California, fueling a growing debate over federal preemption.","Global AI companies must manage model safety, data sources, content labeling, the protection of minors, and high-risk decision-making use cases as separate control items.","Information on AI regulations should be organized into standard fields—such as country/region, effective date, scope of application, obligations, exceptions, penalties, and regulatory agencies—to facilitate searching, auditing, and citation of AI-related information."],"content_markdown":"## Overview: Why Are AI Regulations in July 2026 So “Fragmented”?\n\nAs of July 2026, AI regulations are evolving simultaneously across multiple levels rather than converging into a single global standard. The UN and ITU are establishing a framework for international governance and policy coordination; the EU is refining the risk-based framework of the AI Act into enforceable timelines and procedures; and the U.S. is seeing a rapid proliferation of state-level bills in the absence of unified federal rules.\n\nThis situation can be described as “national fragmentation.” This is because even the same AI model or service is subject to different obligations depending on the region, the intended use, and the target audience. From a corporate perspective, it is not enough to simply “comply with AI regulations”; companies must manage “which provisions apply in each jurisdiction and when they take effect.”\n\n\u003e Reference Date: July 8, 2026. The information below is a summary based on major news reports and official materials from July 2026, as well as known regulatory frameworks, and does not constitute legal advice.\n\n## 1. Key Timeline for July 2026\n\n| Date | Region/Organization | Event | Regulatory Implications |\n|---|---|---|---|\n| 2026-06-29 | EU | Reports indicate that the EU Council has initiated the final approval process for streamlining AI regulations and adjusting the timeline | Impacts AI Act compliance timelines, preparations for high-risk AI implementation, and strategies for utilizing regulatory sandboxes |\n| 2026-07-01 | UN/International | Reports on the UN and ITU’s “AI for Good Global Commission” | A trend aimed at coordinating the international AI governance agenda by connecting corporate CEOs with global political leaders |\n| July 7, 2026 | Illinois, U.S. | Report on the signing of a major AI regulatory bill in Illinois | An example of how state-level AI regulation in the U.S. is effectively driving national operational standards |\n| July 6–7, 2026 | U.S. Federal and State | Coverage of the debate surrounding federal preemption and state regulatory authority | A structural conflict over whether a single federal law should override state laws or allow for state-by-state experimentation |\n| July 8, 2026 | Geneva, Switzerland | First meeting of the UN/ITU AI for Good Global Commission scheduled | A political starting point for coordinating international norms, standards, and policies |\n\n## 2. UN/ITU: A Role Closer to a “Governance Network” Than Law Enforcement\n\nThe UN/ITU’s AI for Good Global Commission is not a mechanism that imposes direct penalties like national laws. Its core purpose is to create a common platform where global political leaders, international organizations, technology companies, and civil society can discuss the risks and opportunities of AI.\n\n### Why Discussions at the UN Level Are Important\n\n- **Cross-border AI Services**: Generative AI, general-purpose models, and cloud-based AI APIs do not operate solely within a single country.\n- **Standardization of Policy Terminology**: There is a need to align terminology internationally regarding safety, reliability, transparency, data sources, bias, and accountability.\n- **Direct Link Between Corporate CEOs and Policymakers**: Technological decisions made by major AI companies can effectively have public policy implications, necessitating coordination channels between governments and businesses.\n- **Participation of Developing and Smaller Nations**: Rules centered solely on the EU, the U.S., and China cannot adequately reflect the interests of users worldwide.\n\n### Practical Implications for Companies\n\nWhile UN discussions do not immediately create compliance obligations, they serve as indicators of the future direction of national legislation and international standards. In particular, the following items recur frequently in discussions at the UN and international organizations, as well as in national laws.\n\n1. Model evaluation and safety testing\n2. Human oversight in high-risk areas\n3. Labeling of AI-generated content\n4. Documentation of data sources and training data\n5. Protection of minors and vulnerable groups\n6. Accountability for AI use in the public sector\n\n## 3. EU: The Significance of the AI Act’s Risk-Based Framework and the 2026 Timeline Adjustment\n\nThe EU AI Act is a representative comprehensive regulation that governs AI systems differently based on their risk level. Its basic structure is divided into prohibited AI practices, high-risk AI, general-purpose AI models, transparency obligations, low-risk AI, and regulatory sandboxes.\n\nA June 29, 2026, EU Council document indicates that the simplification of AI rules, procedural refinements, and schedule adjustments could impact companies’ compliance strategies. While the exact obligations must be confirmed in the final text of the law and the implementing guidelines, from a corporate perspective, the question “What evidence must be prepared by when?” is more important than “Could the timeline be delayed?”\n\n### Areas Companies Should Pay Particular Attention to in the EU AI Act\n\n| Area | Question | Materials to Prepare |\n|---|---|---|\n| High-Risk AI Status | Is it used in sensitive sectors such as hiring, education, credit, insurance, law enforcement, or critical infrastructure? | Classification table of use cases, risk assessment report, review of legal basis |\n| Data Governance | Can the quality and representativeness of training, validation, and test data be explained? | Dataset documentation, bias assessment records, data source documentation |\n| Technical Documentation | Can the design, performance, and limitations of the model or system be explained? | Model Card, System Card, Evaluation Report, Change Log |\n| Human Oversight | Can humans understand and intervene in automated decisions? | Operating Procedures, Administrator Training Materials, Intervention Logs |\n| Transparency | Can users tell when they are interacting with AI or viewing AI-generated content? | Disclosures, UI logs, content source metadata |\n| Sandbox | Is it necessary to conduct experiments with regulatory authorities in a controlled environment? | Test plans, risk mitigation plans, participation application materials |\n\n### Why the EU’s Streamlining Doesn’t Necessarily Mean “Deregulation”\n\nSimplification does not necessarily mean that obligations are eliminated. In fact, it may involve reducing redundant procedures, clarifying timelines and documentation requirements, and increasing access to sandboxes for small and medium-sized enterprises or specific innovation cases. Therefore, rather than halting their compliance preparations, companies should pursue the following three steps in parallel:\n\n- Update the inventory of existing AI systems.\n- Prioritize classifying use cases with high-risk potential.\n- Establish documentation, logging, and evaluation systems that can be reused even if schedules change.\n\n## 4. United States: The Federal Preemption Debate and the Spread of State-Level Regulations\n\nA key characteristic of U.S. AI regulation is the absence of a comprehensive, single federal law. The federal government addresses AI through executive orders, agency-specific guidelines, and existing laws such as consumer protection, anti-discrimination, and competition laws, while state governments are establishing more specific obligations through their own AI legislation.\n\nNews reports on the signing of the Illinois bill in July 2026 and the debate over federal preemption illustrate this structural tension. Federal preemption refers to the principle that federal law takes precedence over state law, thereby limiting or nullifying state regulations. Some in the tech industry may argue that state-by-state rules increase operating costs, while state governments and civil society organizations may argue that state-level safeguards are necessary if federal rules are slow or weak.\n\n### Comparison of Approaches in Major U.S. States\n\n| Category | Illinois | Colorado | California |\n|---|---|---|---|\n| Context as of July 2026 | Reports on the signing of major AI regulatory bills | Known for an approach focused on high-risk AI and preventing algorithmic discrimination | Strong emphasis on generative AI transparency, data and content sourcing, and consumer protection |\n| Regulatory Focus | Potential for comprehensive or stringent state-level AI operational standards | Obligation to ensure AI does not produce discriminatory outcomes in significant decision-making | Labeling of AI-generated content, transparency regarding training data and sources, and platform and consumer protection |\n| Questions for Businesses | Do your operations involve Illinois residents or users of your services within Illinois? | Is AI used for “significant decisions” in areas such as hiring, finance, housing, or education? | Are you required to label generative AI outputs, include watermarks, or indicate sources? |\n| Risk of Fragmentation | If other states adopt similar legislation, it could effectively become a national standard | Definitions of “high-risk AI” and assessment methods may conflict with those of other states | Transparency and labeling requirements directly impact product UI and data pipelines |\n\n### Practical Implications of the Federal Preemption Debate\n\nUntil the federal preemption debate is resolved, companies must assess “state-by-state applicability” rather than “nationwide applicability.” In particular, compliance with state laws must be reviewed in the following cases:\n\n- Providing AI services to residents of a specific state.\n- Using AI for sensitive decision-making in areas such as hiring, credit, insurance, education, housing, and healthcare.\n- Providing or distributing generative AI content to consumers.\n- Operating chatbots, recommendation systems, or educational AI accessible to minors.\n- Providing AI connected to large platforms, advertising, data brokers, employers, or financial institutions.\n\n## 5. Regulatory Checklist for Global AI Companies\n\nAs AI regulations become increasingly fragmented, it is difficult for companies to respond relying solely on their legal teams. Product, data, security, policy, sales, customer support, and public policy teams must share the same regulatory database.\n\n### Minimum Checklist\n\n| Check Item | Description | Example of Responsible Department |\n|---|---|---|\n| Jurisdiction Mapping | Categorize countries, states, language regions, server locations, and user residences where services are provided | Legal, Policy, Data |\n| Use Case Classification | Classify whether AI serves as a simple assistant or influences high-risk decision-making | Product, Legal, Risk |\n| Model Safety | Testing for harmful outputs, hallucinations, security vulnerabilities, and potential for misuse | AI Safety, Security, Quality |\n| Data Sources | Recording the sources and licenses of training, search, and RAG data | Data, Legal |\n| Content Labeling | Methods for notifying users whether content is AI-generated or modified | Product, Design, Policy |\n| Protection of Minors | Age verification, safety filters, parental controls, restrictions on sensitive conversations | Trust \u0026 Safety, Product |\n| High-Risk Use Controls | Restrictions or separate approvals for sensitive use cases such as hiring, credit, healthcare, education, and law enforcement | Sales, Legal, Compliance |\n| Audit Logs | Records of model versions, prompts, outputs, user actions, and error responses | Engineering, Security |\n| Supply Chain Management | Mandatory verification of external models, APIs, data providers, and plugins | Procurement, Security, Legal |\n| Incident Response | Reporting and investigation procedures in the event of an AI incident or inquiry from regulatory authorities | Security, Legal, PR |\n\n## 6. Standard Fields Required for Expansion into Data Articles\n\nAI regulations change rapidly, making it difficult to manage them using narrative articles alone. To make it easier for search engines and AI systems to reference them, each rule must be organized into standard data fields.\n\n| Field Name | Description | Example Values |\n|---|---|---|\n| jurisdiction | Country, region, state, international organization | EU, US-IL, US-CO, UN/ITU |\n| instrument_type | Law, regulation, guideline, commission, executive order | Regulation, State Act, Commission |\n| status | Proposed, passed, signed, in force, under revision | signed, approved, in force |\n| adoption_date | Date of adoption or signing | 2026-07-07 |\n| effective_date | Effective or implementation date | To be confirmed |\n| regulated_entities | Developers, distributors, deployers, platforms, public institutions, etc. | AI deployers, developers |\n| covered_systems | General-purpose AI, high-risk AI, generative AI, automated decision-making, etc. | high-risk AI systems |\n| core_obligations | Key obligations | risk assessment, transparency, documentation |\n| exemptions | Exceptions | Research, open source, small businesses, etc. (to be confirmed) |\n| penalties | Penalties or enforcement measures | administrative fines, civil enforcement, orders from supervisory authorities |\n| regulator | Regulatory bodies | EU AI Office, state attorney general, etc. |\n| source_url | URL of official document or reliable news report | Link to original text |\n| last_reviewed | Last review date | 2026-07-08 |\n\nThis field-based structuring is important not only for regulatory compliance but also for AI search and citation. A machine-readable structure—such as “US-IL, 2026-07-07, signed, high-risk AI, transparency obligations”—yields more accurate search results than broad terms like “U.S. AI regulation.”\n\n## 7. Practical Interpretation: A Single Global Policy and Regional Annexes Are Needed\n\nGlobal AI companies cannot create completely different products for every region. Conversely, it is also difficult to comply with all regional regulations using a single internal policy. A practical approach is as follows:\n\n1. Establish **company-wide AI principles**: safety, transparency, accountability, privacy protection, and anti-discrimination.\n2. Establish **region-specific annexes**: EU high-risk AI, state-by-state automated decision-making in the U.S., California-style content labeling, Illinois-style obligations, etc.\n3. Implement **product launch gates**: New AI features must be reviewed for jurisdiction, risk, data, and labeling requirements before launch.\n4. Create an **evidence-based compliance system**: Actual test results, logs, training records, and change histories are more important than policy documents.\n5. **Digitize regulatory change monitoring**: Update press releases, legal texts, and regulatory agency guidelines using standardized fields.\n\n## Conclusion\n\nThe AI regulatory landscape in July 2026 is a multi-layered structure resulting from the convergence of international coordination by the UN, risk-based enforcement of legislation by the EU, and the proliferation of individual state laws in the U.S. Rather than waiting to see “which regulation will ultimately prevail,” companies must manage obligations by jurisdiction as if they were data and control high-risk use cases, transparency, model safety, and data sources through repeatable procedures.\n\nWhile the fragmentation of AI regulations increases costs in the short term, well-structured compliance data and internal governance can become a competitive advantage for trustworthy AI products in the long run.","content_html":"\u003ch2\u003e\u003ca href=\"#overview-why-are-ai-regulations-in-july-2026-so-fragmented\" class=\"anchor\" id=\"overview-why-are-ai-regulations-in-july-2026-so-fragmented\"\u003e\u003c/a\u003eOverview: Why Are AI Regulations in July 2026 So “Fragmented”?\u003c/h2\u003e\n\u003cp\u003eAs of July 2026, AI regulations are evolving simultaneously across multiple levels rather than converging into a single global standard. The UN and ITU are establishing a framework for international governance and policy coordination; the EU is refining the risk-based framework of the AI Act into enforceable timelines and procedures; and the U.S. is seeing a rapid proliferation of state-level bills in the absence of unified federal rules.\u003c/p\u003e\n\u003cp\u003eThis situation can be described as “national fragmentation.” This is because even the same AI model or service is subject to different obligations depending on the region, the intended use, and the target audience. From a corporate perspective, it is not enough to simply “comply with AI regulations”; companies must manage “which provisions apply in each jurisdiction and when they take effect.”\u003c/p\u003e\n\u003cblockquote\u003e\n\u003cp\u003eReference Date: July 8, 2026. The information below is a summary based on major news reports and official materials from July 2026, as well as known regulatory frameworks, and does not constitute legal advice.\u003c/p\u003e\n\u003c/blockquote\u003e\n\u003ch2\u003e\u003ca href=\"#1-key-timeline-for-july-2026\" class=\"anchor\" id=\"1-key-timeline-for-july-2026\"\u003e\u003c/a\u003e1. Key Timeline for July 2026\u003c/h2\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eDate\u003c/th\u003e\n\u003cth\u003eRegion/Organization\u003c/th\u003e\n\u003cth\u003eEvent\u003c/th\u003e\n\u003cth\u003eRegulatory Implications\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003e2026-06-29\u003c/td\u003e\n\u003ctd\u003eEU\u003c/td\u003e\n\u003ctd\u003eReports indicate that the EU Council has initiated the final approval process for streamlining AI regulations and adjusting the timeline\u003c/td\u003e\n\u003ctd\u003eImpacts AI Act compliance timelines, preparations for high-risk AI implementation, and strategies for utilizing regulatory sandboxes\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003e2026-07-01\u003c/td\u003e\n\u003ctd\u003eUN/International\u003c/td\u003e\n\u003ctd\u003eReports on the UN and ITU’s “AI for Good Global Commission”\u003c/td\u003e\n\u003ctd\u003eA trend aimed at coordinating the international AI governance agenda by connecting corporate CEOs with global political leaders\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eJuly 7, 2026\u003c/td\u003e\n\u003ctd\u003eIllinois, U.S.\u003c/td\u003e\n\u003ctd\u003eReport on the signing of a major AI regulatory bill in Illinois\u003c/td\u003e\n\u003ctd\u003eAn example of how state-level AI regulation in the U.S. is effectively driving national operational standards\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eJuly 6–7, 2026\u003c/td\u003e\n\u003ctd\u003eU.S. Federal and State\u003c/td\u003e\n\u003ctd\u003eCoverage of the debate surrounding federal preemption and state regulatory authority\u003c/td\u003e\n\u003ctd\u003eA structural conflict over whether a single federal law should override state laws or allow for state-by-state experimentation\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eJuly 8, 2026\u003c/td\u003e\n\u003ctd\u003eGeneva, Switzerland\u003c/td\u003e\n\u003ctd\u003eFirst meeting of the UN/ITU AI for Good Global Commission scheduled\u003c/td\u003e\n\u003ctd\u003eA political starting point for coordinating international norms, standards, and policies\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003ch2\u003e\u003ca href=\"#2-unitu-a-role-closer-to-a-governance-network-than-law-enforcement\" class=\"anchor\" id=\"2-unitu-a-role-closer-to-a-governance-network-than-law-enforcement\"\u003e\u003c/a\u003e2. UN/ITU: A Role Closer to a “Governance Network” Than Law Enforcement\u003c/h2\u003e\n\u003cp\u003eThe UN/ITU’s AI for Good Global Commission is not a mechanism that imposes direct penalties like national laws. Its core purpose is to create a common platform where global political leaders, international organizations, technology companies, and civil society can discuss the risks and opportunities of AI.\u003c/p\u003e\n\u003ch3\u003e\u003ca href=\"#why-discussions-at-the-un-level-are-important\" class=\"anchor\" id=\"why-discussions-at-the-un-level-are-important\"\u003e\u003c/a\u003eWhy Discussions at the UN Level Are Important\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eCross-border AI Services\u003c/strong\u003e: Generative AI, general-purpose models, and cloud-based AI APIs do not operate solely within a single country.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eStandardization of Policy Terminology\u003c/strong\u003e: There is a need to align terminology internationally regarding safety, reliability, transparency, data sources, bias, and accountability.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eDirect Link Between Corporate CEOs and Policymakers\u003c/strong\u003e: Technological decisions made by major AI companies can effectively have public policy implications, necessitating coordination channels between governments and businesses.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eParticipation of Developing and Smaller Nations\u003c/strong\u003e: Rules centered solely on the EU, the U.S., and China cannot adequately reflect the interests of users worldwide.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3\u003e\u003ca href=\"#practical-implications-for-companies\" class=\"anchor\" id=\"practical-implications-for-companies\"\u003e\u003c/a\u003ePractical Implications for Companies\u003c/h3\u003e\n\u003cp\u003eWhile UN discussions do not immediately create compliance obligations, they serve as indicators of the future direction of national legislation and international standards. In particular, the following items recur frequently in discussions at the UN and international organizations, as well as in national laws.\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003eModel evaluation and safety testing\u003c/li\u003e\n\u003cli\u003eHuman oversight in high-risk areas\u003c/li\u003e\n\u003cli\u003eLabeling of AI-generated content\u003c/li\u003e\n\u003cli\u003eDocumentation of data sources and training data\u003c/li\u003e\n\u003cli\u003eProtection of minors and vulnerable groups\u003c/li\u003e\n\u003cli\u003eAccountability for AI use in the public sector\u003c/li\u003e\n\u003c/ol\u003e\n\u003ch2\u003e\u003ca href=\"#3-eu-the-significance-of-the-ai-acts-risk-based-framework-and-the-2026-timeline-adjustment\" class=\"anchor\" id=\"3-eu-the-significance-of-the-ai-acts-risk-based-framework-and-the-2026-timeline-adjustment\"\u003e\u003c/a\u003e3. EU: The Significance of the AI Act’s Risk-Based Framework and the 2026 Timeline Adjustment\u003c/h2\u003e\n\u003cp\u003eThe EU AI Act is a representative comprehensive regulation that governs AI systems differently based on their risk level. Its basic structure is divided into prohibited AI practices, high-risk AI, general-purpose AI models, transparency obligations, low-risk AI, and regulatory sandboxes.\u003c/p\u003e\n\u003cp\u003eA June 29, 2026, EU Council document indicates that the simplification of AI rules, procedural refinements, and schedule adjustments could impact companies’ compliance strategies. While the exact obligations must be confirmed in the final text of the law and the implementing guidelines, from a corporate perspective, the question “What evidence must be prepared by when?” is more important than “Could the timeline be delayed?”\u003c/p\u003e\n\u003ch3\u003e\u003ca href=\"#areas-companies-should-pay-particular-attention-to-in-the-eu-ai-act\" class=\"anchor\" id=\"areas-companies-should-pay-particular-attention-to-in-the-eu-ai-act\"\u003e\u003c/a\u003eAreas Companies Should Pay Particular Attention to in the EU AI Act\u003c/h3\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eArea\u003c/th\u003e\n\u003cth\u003eQuestion\u003c/th\u003e\n\u003cth\u003eMaterials to Prepare\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eHigh-Risk AI Status\u003c/td\u003e\n\u003ctd\u003eIs it used in sensitive sectors such as hiring, education, credit, insurance, law enforcement, or critical infrastructure?\u003c/td\u003e\n\u003ctd\u003eClassification table of use cases, risk assessment report, review of legal basis\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eData Governance\u003c/td\u003e\n\u003ctd\u003eCan the quality and representativeness of training, validation, and test data be explained?\u003c/td\u003e\n\u003ctd\u003eDataset documentation, bias assessment records, data source documentation\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eTechnical Documentation\u003c/td\u003e\n\u003ctd\u003eCan the design, performance, and limitations of the model or system be explained?\u003c/td\u003e\n\u003ctd\u003eModel Card, System Card, Evaluation Report, Change Log\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eHuman Oversight\u003c/td\u003e\n\u003ctd\u003eCan humans understand and intervene in automated decisions?\u003c/td\u003e\n\u003ctd\u003eOperating Procedures, Administrator Training Materials, Intervention Logs\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eTransparency\u003c/td\u003e\n\u003ctd\u003eCan users tell when they are interacting with AI or viewing AI-generated content?\u003c/td\u003e\n\u003ctd\u003eDisclosures, UI logs, content source metadata\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eSandbox\u003c/td\u003e\n\u003ctd\u003eIs it necessary to conduct experiments with regulatory authorities in a controlled environment?\u003c/td\u003e\n\u003ctd\u003eTest plans, risk mitigation plans, participation application materials\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003ch3\u003e\u003ca href=\"#why-the-eus-streamlining-doesnt-necessarily-mean-deregulation\" class=\"anchor\" id=\"why-the-eus-streamlining-doesnt-necessarily-mean-deregulation\"\u003e\u003c/a\u003eWhy the EU’s Streamlining Doesn’t Necessarily Mean “Deregulation”\u003c/h3\u003e\n\u003cp\u003eSimplification does not necessarily mean that obligations are eliminated. In fact, it may involve reducing redundant procedures, clarifying timelines and documentation requirements, and increasing access to sandboxes for small and medium-sized enterprises or specific innovation cases. Therefore, rather than halting their compliance preparations, companies should pursue the following three steps in parallel:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eUpdate the inventory of existing AI systems.\u003c/li\u003e\n\u003cli\u003ePrioritize classifying use cases with high-risk potential.\u003c/li\u003e\n\u003cli\u003eEstablish documentation, logging, and evaluation systems that can be reused even if schedules change.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003e\u003ca href=\"#4-united-states-the-federal-preemption-debate-and-the-spread-of-state-level-regulations\" class=\"anchor\" id=\"4-united-states-the-federal-preemption-debate-and-the-spread-of-state-level-regulations\"\u003e\u003c/a\u003e4. United States: The Federal Preemption Debate and the Spread of State-Level Regulations\u003c/h2\u003e\n\u003cp\u003eA key characteristic of U.S. AI regulation is the absence of a comprehensive, single federal law. The federal government addresses AI through executive orders, agency-specific guidelines, and existing laws such as consumer protection, anti-discrimination, and competition laws, while state governments are establishing more specific obligations through their own AI legislation.\u003c/p\u003e\n\u003cp\u003eNews reports on the signing of the Illinois bill in July 2026 and the debate over federal preemption illustrate this structural tension. Federal preemption refers to the principle that federal law takes precedence over state law, thereby limiting or nullifying state regulations. Some in the tech industry may argue that state-by-state rules increase operating costs, while state governments and civil society organizations may argue that state-level safeguards are necessary if federal rules are slow or weak.\u003c/p\u003e\n\u003ch3\u003e\u003ca href=\"#comparison-of-approaches-in-major-us-states\" class=\"anchor\" id=\"comparison-of-approaches-in-major-us-states\"\u003e\u003c/a\u003eComparison of Approaches in Major U.S. States\u003c/h3\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eCategory\u003c/th\u003e\n\u003cth\u003eIllinois\u003c/th\u003e\n\u003cth\u003eColorado\u003c/th\u003e\n\u003cth\u003eCalifornia\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eContext as of July 2026\u003c/td\u003e\n\u003ctd\u003eReports on the signing of major AI regulatory bills\u003c/td\u003e\n\u003ctd\u003eKnown for an approach focused on high-risk AI and preventing algorithmic discrimination\u003c/td\u003e\n\u003ctd\u003eStrong emphasis on generative AI transparency, data and content sourcing, and consumer protection\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eRegulatory Focus\u003c/td\u003e\n\u003ctd\u003ePotential for comprehensive or stringent state-level AI operational standards\u003c/td\u003e\n\u003ctd\u003eObligation to ensure AI does not produce discriminatory outcomes in significant decision-making\u003c/td\u003e\n\u003ctd\u003eLabeling of AI-generated content, transparency regarding training data and sources, and platform and consumer protection\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eQuestions for Businesses\u003c/td\u003e\n\u003ctd\u003eDo your operations involve Illinois residents or users of your services within Illinois?\u003c/td\u003e\n\u003ctd\u003eIs AI used for “significant decisions” in areas such as hiring, finance, housing, or education?\u003c/td\u003e\n\u003ctd\u003eAre you required to label generative AI outputs, include watermarks, or indicate sources?\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eRisk of Fragmentation\u003c/td\u003e\n\u003ctd\u003eIf other states adopt similar legislation, it could effectively become a national standard\u003c/td\u003e\n\u003ctd\u003eDefinitions of “high-risk AI” and assessment methods may conflict with those of other states\u003c/td\u003e\n\u003ctd\u003eTransparency and labeling requirements directly impact product UI and data pipelines\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003ch3\u003e\u003ca href=\"#practical-implications-of-the-federal-preemption-debate\" class=\"anchor\" id=\"practical-implications-of-the-federal-preemption-debate\"\u003e\u003c/a\u003ePractical Implications of the Federal Preemption Debate\u003c/h3\u003e\n\u003cp\u003eUntil the federal preemption debate is resolved, companies must assess “state-by-state applicability” rather than “nationwide applicability.” In particular, compliance with state laws must be reviewed in the following cases:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eProviding AI services to residents of a specific state.\u003c/li\u003e\n\u003cli\u003eUsing AI for sensitive decision-making in areas such as hiring, credit, insurance, education, housing, and healthcare.\u003c/li\u003e\n\u003cli\u003eProviding or distributing generative AI content to consumers.\u003c/li\u003e\n\u003cli\u003eOperating chatbots, recommendation systems, or educational AI accessible to minors.\u003c/li\u003e\n\u003cli\u003eProviding AI connected to large platforms, advertising, data brokers, employers, or financial institutions.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003e\u003ca href=\"#5-regulatory-checklist-for-global-ai-companies\" class=\"anchor\" id=\"5-regulatory-checklist-for-global-ai-companies\"\u003e\u003c/a\u003e5. Regulatory Checklist for Global AI Companies\u003c/h2\u003e\n\u003cp\u003eAs AI regulations become increasingly fragmented, it is difficult for companies to respond relying solely on their legal teams. Product, data, security, policy, sales, customer support, and public policy teams must share the same regulatory database.\u003c/p\u003e\n\u003ch3\u003e\u003ca href=\"#minimum-checklist\" class=\"anchor\" id=\"minimum-checklist\"\u003e\u003c/a\u003eMinimum Checklist\u003c/h3\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eCheck Item\u003c/th\u003e\n\u003cth\u003eDescription\u003c/th\u003e\n\u003cth\u003eExample of Responsible Department\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eJurisdiction Mapping\u003c/td\u003e\n\u003ctd\u003eCategorize countries, states, language regions, server locations, and user residences where services are provided\u003c/td\u003e\n\u003ctd\u003eLegal, Policy, Data\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eUse Case Classification\u003c/td\u003e\n\u003ctd\u003eClassify whether AI serves as a simple assistant or influences high-risk decision-making\u003c/td\u003e\n\u003ctd\u003eProduct, Legal, Risk\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eModel Safety\u003c/td\u003e\n\u003ctd\u003eTesting for harmful outputs, hallucinations, security vulnerabilities, and potential for misuse\u003c/td\u003e\n\u003ctd\u003eAI Safety, Security, Quality\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eData Sources\u003c/td\u003e\n\u003ctd\u003eRecording the sources and licenses of training, search, and RAG data\u003c/td\u003e\n\u003ctd\u003eData, Legal\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eContent Labeling\u003c/td\u003e\n\u003ctd\u003eMethods for notifying users whether content is AI-generated or modified\u003c/td\u003e\n\u003ctd\u003eProduct, Design, Policy\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eProtection of Minors\u003c/td\u003e\n\u003ctd\u003eAge verification, safety filters, parental controls, restrictions on sensitive conversations\u003c/td\u003e\n\u003ctd\u003eTrust \u0026amp; Safety, Product\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eHigh-Risk Use Controls\u003c/td\u003e\n\u003ctd\u003eRestrictions or separate approvals for sensitive use cases such as hiring, credit, healthcare, education, and law enforcement\u003c/td\u003e\n\u003ctd\u003eSales, Legal, Compliance\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eAudit Logs\u003c/td\u003e\n\u003ctd\u003eRecords of model versions, prompts, outputs, user actions, and error responses\u003c/td\u003e\n\u003ctd\u003eEngineering, Security\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eSupply Chain Management\u003c/td\u003e\n\u003ctd\u003eMandatory verification of external models, APIs, data providers, and plugins\u003c/td\u003e\n\u003ctd\u003eProcurement, Security, Legal\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eIncident Response\u003c/td\u003e\n\u003ctd\u003eReporting and investigation procedures in the event of an AI incident or inquiry from regulatory authorities\u003c/td\u003e\n\u003ctd\u003eSecurity, Legal, PR\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003ch2\u003e\u003ca href=\"#6-standard-fields-required-for-expansion-into-data-articles\" class=\"anchor\" id=\"6-standard-fields-required-for-expansion-into-data-articles\"\u003e\u003c/a\u003e6. Standard Fields Required for Expansion into Data Articles\u003c/h2\u003e\n\u003cp\u003eAI regulations change rapidly, making it difficult to manage them using narrative articles alone. To make it easier for search engines and AI systems to reference them, each rule must be organized into standard data fields.\u003c/p\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eField Name\u003c/th\u003e\n\u003cth\u003eDescription\u003c/th\u003e\n\u003cth\u003eExample Values\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003ejurisdiction\u003c/td\u003e\n\u003ctd\u003eCountry, region, state, international organization\u003c/td\u003e\n\u003ctd\u003eEU, US-IL, US-CO, UN/ITU\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003einstrument_type\u003c/td\u003e\n\u003ctd\u003eLaw, regulation, guideline, commission, executive order\u003c/td\u003e\n\u003ctd\u003eRegulation, State Act, Commission\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003estatus\u003c/td\u003e\n\u003ctd\u003eProposed, passed, signed, in force, under revision\u003c/td\u003e\n\u003ctd\u003esigned, approved, in force\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eadoption_date\u003c/td\u003e\n\u003ctd\u003eDate of adoption or signing\u003c/td\u003e\n\u003ctd\u003e2026-07-07\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eeffective_date\u003c/td\u003e\n\u003ctd\u003eEffective or implementation date\u003c/td\u003e\n\u003ctd\u003eTo be confirmed\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eregulated_entities\u003c/td\u003e\n\u003ctd\u003eDevelopers, distributors, deployers, platforms, public institutions, etc.\u003c/td\u003e\n\u003ctd\u003eAI deployers, developers\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ecovered_systems\u003c/td\u003e\n\u003ctd\u003eGeneral-purpose AI, high-risk AI, generative AI, automated decision-making, etc.\u003c/td\u003e\n\u003ctd\u003ehigh-risk AI systems\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ecore_obligations\u003c/td\u003e\n\u003ctd\u003eKey obligations\u003c/td\u003e\n\u003ctd\u003erisk assessment, transparency, documentation\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eexemptions\u003c/td\u003e\n\u003ctd\u003eExceptions\u003c/td\u003e\n\u003ctd\u003eResearch, open source, small businesses, etc. (to be confirmed)\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003epenalties\u003c/td\u003e\n\u003ctd\u003ePenalties or enforcement measures\u003c/td\u003e\n\u003ctd\u003eadministrative fines, civil enforcement, orders from supervisory authorities\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eregulator\u003c/td\u003e\n\u003ctd\u003eRegulatory bodies\u003c/td\u003e\n\u003ctd\u003eEU AI Office, state attorney general, etc.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003esource_url\u003c/td\u003e\n\u003ctd\u003eURL of official document or reliable news report\u003c/td\u003e\n\u003ctd\u003eLink to original text\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003elast_reviewed\u003c/td\u003e\n\u003ctd\u003eLast review date\u003c/td\u003e\n\u003ctd\u003e2026-07-08\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003cp\u003eThis field-based structuring is important not only for regulatory compliance but also for AI search and citation. A machine-readable structure—such as “US-IL, 2026-07-07, signed, high-risk AI, transparency obligations”—yields more accurate search results than broad terms like “U.S. AI regulation.”\u003c/p\u003e\n\u003ch2\u003e\u003ca href=\"#7-practical-interpretation-a-single-global-policy-and-regional-annexes-are-needed\" class=\"anchor\" id=\"7-practical-interpretation-a-single-global-policy-and-regional-annexes-are-needed\"\u003e\u003c/a\u003e7. Practical Interpretation: A Single Global Policy and Regional Annexes Are Needed\u003c/h2\u003e\n\u003cp\u003eGlobal AI companies cannot create completely different products for every region. Conversely, it is also difficult to comply with all regional regulations using a single internal policy. A practical approach is as follows:\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003eEstablish \u003cstrong\u003ecompany-wide AI principles\u003c/strong\u003e: safety, transparency, accountability, privacy protection, and anti-discrimination.\u003c/li\u003e\n\u003cli\u003eEstablish \u003cstrong\u003eregion-specific annexes\u003c/strong\u003e: EU high-risk AI, state-by-state automated decision-making in the U.S., California-style content labeling, Illinois-style obligations, etc.\u003c/li\u003e\n\u003cli\u003eImplement \u003cstrong\u003eproduct launch gates\u003c/strong\u003e: New AI features must be reviewed for jurisdiction, risk, data, and labeling requirements before launch.\u003c/li\u003e\n\u003cli\u003eCreate an \u003cstrong\u003eevidence-based compliance system\u003c/strong\u003e: Actual test results, logs, training records, and change histories are more important than policy documents.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eDigitize regulatory change monitoring\u003c/strong\u003e: Update press releases, legal texts, and regulatory agency guidelines using standardized fields.\u003c/li\u003e\n\u003c/ol\u003e\n\u003ch2\u003e\u003ca href=\"#conclusion\" class=\"anchor\" id=\"conclusion\"\u003e\u003c/a\u003eConclusion\u003c/h2\u003e\n\u003cp\u003eThe AI regulatory landscape in July 2026 is a multi-layered structure resulting from the convergence of international coordination by the UN, risk-based enforcement of legislation by the EU, and the proliferation of individual state laws in the U.S. Rather than waiting to see “which regulation will ultimately prevail,” companies must manage obligations by jurisdiction as if they were data and control high-risk use cases, transparency, model safety, and data sources through repeatable procedures.\u003c/p\u003e\n\u003cp\u003eWhile the fragmentation of AI regulations increases costs in the short term, well-structured compliance data and internal governance can become a competitive advantage for trustworthy AI products in the long run.\u003c/p\u003e\n","tags":["AI","UN","EU AI Act","US state law","Governance"],"faqs":[{"question":"What Is “National Fragmentation” in AI Regulation?","answer":"This refers to the phenomenon where, even for the same AI technology, different obligations and effective dates apply depending on the country, region, state, or use case. For example, the EU is implementing the Risk-Based AI Act, the United States is seeing an increase in state-level legislation, and the UN is coordinating discussions on international governance."},{"question":"Is the UN/ITU’s AI for Good Global Commission legally binding?","answer":"Generally speaking, UN and ITU committees themselves are not enforcement agencies that directly impose fines, as national laws do. However, they can influence international agenda-setting, policy coordination, and the formation of consultation frameworks between companies and governments."},{"question":"What is the first thing companies need to do under the EU AI Act?","answer":"You must create a list of your company’s AI systems and classify each system as either high-risk AI, general-purpose AI, generative AI, or low-risk AI. You must then prepare information on data sources, risk assessments, technical documentation, human oversight, and transparency disclosures."},{"question":"Does the simplification of the EU's AI regulations mean it's possible to avoid regulation?","answer":"It cannot be concluded so definitively. Simplification may refer to reducing redundant procedures or adjusting schedule and documentation requirements, while core obligations may still remain in place. Companies should review the final text of the law and the regulatory agency’s guidelines."},{"question":"Why is the debate over federal preemption in the United States important?","answer":"Federal preemption is a debate over whether federal law takes precedence over state law, thereby limiting state-level AI regulations. If preemption is strictly applied, state-level regulations may decrease; otherwise, regulations in states like Illinois, Colorado, and California may continue to expand."},{"question":"What is the key to Colorado's approach to AI regulation?","answer":"Colorado is known for its approach aimed at reducing algorithmic bias that may arise from high-risk AI and automated high-stakes decision-making. Companies must assess the impact of AI in sensitive areas such as hiring, finance, housing, and education."},{"question":"How does the California approach differ from that of other states?","answer":"California is at the forefront of discussions regarding the labeling of generative AI content, attribution and transparency, consumer protection, and platform liability. Therefore, product UI, watermarks, content metadata, and descriptions of training data may become key compliance requirements."},{"question":"What are the key areas that global AI companies need to manage collectively?","answer":"Key elements include jurisdiction mapping, risk classification by use case, model safety testing, documentation of data sources, labeling of AI-generated content, protection of minors, restrictions on high-risk uses, audit logs, supply chain management, and incident response procedures."},{"question":"What are the benefits of standardizing AI regulatory data?","answer":"Managing standard fields such as country/region, effective date, scope of application, obligations, exceptions, penalties, and regulatory agencies is beneficial for legal reviews, product launch reviews, audit responses, search engine visibility, and accurate citations by AI systems."}],"sources":[{"url":"https://www.axios.com/2026/07/01/un-ai-commission-ceos-world-leaders","title":"Axios - UN AI Commission Brings Together CEOs and World Leaders","type":"source"},{"url":"https://www.consilium.europa.eu/en/press/press-releases/2026/06/29/artificial-intelligence-council-gives-final-green-light-to-simplify-and-streamline-rules/pdf/","title":"Council of the European Union - Artificial Intelligence: Council Gives Final Approval to Simplify and Streamline Rules","type":"source"},{"url":"https://www.washingtonpost.com/business/2026/07/07/landmark-ai-regulations-illinois-statedriven-national-standard/b046234a-7a29-11f1-b194-f872dd4ec5aa_story.html","title":"The Washington Post - Landmark AI regulations in Illinois and a state-led national standard","type":"source"},{"url":"https://www.insideglobaltech.com/2026/07/06/backlash-to-bipartisan-ai-omnibus-illustrates-preemption-impasse/","title":"Inside Global Tech - Backlash to bipartisan AI omnibus highlights impasse over preemption","type":"source"}],"images":[{"id":83,"url":"https://injoys.com/rails/active_storage/blobs/redirect/eyJfcmFpbHMiOnsiZGF0YSI6ODAwLCJwdXIiOiJibG9iX2lkIn19--60da7526e12a8e7247d5a693b68f4c8ba13a9ace/ai-0e503730.webp","is_representative":true,"generation_method":"ai_image","license":"ai_generated","mime_type":"image/webp","translations":{"ko":{"alt":"퍼즐 조각 세계지도와 중앙 AI 칩, UN 회의, EU 문서, 미국 주 지도","caption":"세계와 지역별 AI 규제 체계가 UN, EU, 미국 주 단위로 나뉘어 표시되어 있다.","description":null},"en":{"alt":"Puzzle-piece world map with a central AI chip, UN meeting, EU files, and U.S. state map","caption":"The illustration shows fragmented AI governance across the UN, EU, and U.S. states.","description":null},"ja":{"alt":"パズル状の世界地図、中央のAIチップ、国連会議、EU文書、米国州地図","caption":"国連、EU、米国各州に分かれるAI規制の広がりを示している。","description":null},"es":{"alt":"Mapa mundial en piezas de rompecabezas con chip de IA, reunión de la ONU, archivos de la UE y mapa de EE. UU.","caption":"La ilustración muestra la fragmentación de la gobernanza de la IA entre la ONU, la UE y los estados de EE. UU.","description":null},"id":{"alt":"Peta dunia berbentuk kepingan puzzle dengan chip AI, rapat PBB, dokumen UE, dan peta negara bagian AS","caption":"Ilustrasi ini menunjukkan tata kelola AI yang terpecah antara PBB, UE, dan negara bagian AS.","description":null},"pt":{"alt":"Mapa-múndi em peças de quebra-cabeça com chip de IA, reunião da ONU, arquivos da UE e mapa dos EUA","caption":"A ilustração mostra a fragmentação da governança de IA entre ONU, UE e estados dos EUA.","description":null},"zh-hant":{"alt":"拼圖式世界地圖搭配中央 AI 晶片、聯合國會議、歐盟文件與美國州地圖","caption":"插圖呈現聯合國、歐盟與美國各州之間分散的 AI 治理格局。","description":null}}},{"id":84,"url":"https://injoys.com/rails/active_storage/blobs/redirect/eyJfcmFpbHMiOnsiZGF0YSI6ODA2LCJwdXIiOiJibG9iX2lkIn19--b85e630037b2faad2796009eac522381e6b2cf37/ai-93d3a795.webp","is_representative":false,"generation_method":"ai_image","license":"ai_generated","mime_type":"image/webp","translations":{"ko":{"alt":"지구본과 AI 규제 대시보드, UN·EU·미국 주별 규칙을 보여주는 인포그래픽","caption":"글로벌 AI 규제가 UN, EU, 미국 주 단위로 나뉘어 표시된 개념도입니다.","description":null},"en":{"alt":"Global AI regulation dashboard with UN, EU, and U.S. state rule panels","caption":"The infographic shows AI rules fragmented across the UN, EU, and U.S. states.","description":null},"ja":{"alt":"地球儀とAI規制ダッシュボード、国連・EU・米州規則のパネル","caption":"AI規制が国連、EU、米国の州ごとに分かれて示されています。","description":null},"es":{"alt":"Panel global de regulación de IA con secciones de la ONU, la UE y estados de EE. UU.","caption":"La infografía muestra reglas de IA fragmentadas entre la ONU, la UE y los estados de EE. UU.","description":null},"id":{"alt":"Dasbor regulasi AI global dengan panel PBB, Uni Eropa, dan aturan negara bagian AS","caption":"Infografik ini menampilkan aturan AI yang terpecah di PBB, Uni Eropa, dan negara bagian AS.","description":null},"pt":{"alt":"Painel global de regulação de IA com ONU, UE e regras de estados dos EUA","caption":"O infográfico mostra regras de IA fragmentadas entre a ONU, a UE e estados dos EUA.","description":null},"zh-hant":{"alt":"全球 AI 監管儀表板，包含聯合國、歐盟與美國州級規則面板","caption":"這張資訊圖呈現 AI 規則在聯合國、歐盟與美國各州之間的分散情況。","description":null}}}],"published_at":"2026-07-08T10:14:39+09:00","updated_at":"2026-07-08T10:14:39+09:00","license":"cc_by","translation_status":"reviewed","available_locales":["ko","en","ja","es"],"data_locales":["ko","en","ja","es","id","pt","zh-hant"],"url":"https://injoys.com/en/articles/ai-regulation-fragmentation-un-eu-us-states-july-2026"}