A timeline of major regulations, laws, and policy actions shaping the future of artificial intelligence worldwide.
The White House published a National Policy Framework for Artificial Intelligence outlining seven guiding principles for federal AI legislation. The framework calls for protecting children online, safeguarding communities, respecting intellectual property, preventing censorship, enabling innovation, developing an AI-ready workforce, and establishing federal standards that would pre-empt state-level AI laws.
The framework drew bipartisan attention but also opposition. Democrats introduced the GUARDRAILS Act to prevent federal pre-emption of state AI protections. The proposal's recommendation against creating a new federal AI regulatory body signals a preference for sector-specific oversight through existing agencies.
Read more →The EU Council agreed to streamline rules on artificial intelligence by extending the deadline for high-risk AI system compliance by up to 16 months. The revised timeline ensures rules take effect only after the Commission confirms that the necessary technical standards and compliance tools are available for providers.
The extension acknowledges that enforcement infrastructure across member states remains incomplete, with only eight of 27 countries having designated competent authorities by the original deadline. The adjustment aims to balance regulatory ambition with practical readiness.
Read more →South Korea's AI Basic Act, introduced in January 2025, enters broader enforcement with mandatory requirements for high-risk AI systems. The law establishes national AI governance standards and transparency obligations for major AI providers.
South Korea joins the EU as a major jurisdiction with comprehensive AI legislation, representing the global trend toward proactive governance.
Read more →Taiwan enacts its AI Basic Act, establishing a human-centric framework for artificial intelligence governance. The law designates the National Science and Technology Council (NSTC) as the central authority and establishes four core principles: human-centric development, sustainable development, effective governance, and reasonable accountability.
Taiwan joins South Korea and the EU in establishing comprehensive AI legislation, contributing to Asia's growing regulatory infrastructure for AI systems.
Source: Taiwan NSTC →New York Governor Kathy Hochul signs the RAISE Act (Responsible Artificial Intelligence Systems Evaluation Act), making New York the second US state to enact comprehensive frontier AI regulation. The law requires AI developers working on frontier models to implement safety protocols and conduct impact assessments, with mandatory 72-hour incident reporting to the state.
The law establishes civil penalties of up to $1 million for first offenses and $3 million for subsequent violations, positioning New York alongside Colorado and California in state-level AI regulation efforts.
Source: New York Governor's Office →An executive order establishes a federal AI policy framework and creates the AI Litigation Task Force within the Department of Justice. The order asserts federal preemption over state-level AI regulations and directs the task force to challenge state laws deemed to conflict with federal policy.
The action reflects ongoing tension between federal and state regulatory approaches, with significant implications for how state-level AI laws like those in Colorado, California, and New York will be implemented and enforced.
Source: White House / White & Case Analysis →A Munich court issues a landmark ruling in a case brought by GEMA (German music rights organization) against OpenAI. The court finds that ChatGPT reproducing copyrighted song lyrics violates German law. The court orders OpenAI to refrain from reproducing copyrighted lyrics and to compensate affected rights holders.
The ruling sets an important precedent for AI copyright liability in Europe and supports arguments that training data and output must respect existing copyright protections.
Source: Euronews →California Governor Newsom signs Senate Bill 53 (SB-53), the Transparency in Frontier AI Act. The law requires developers of frontier AI models (defined as those trained on 10^26 or more compute) to publish comprehensive transparency reports before deploying systems in California. The law takes effect January 1, 2026.
California advances its position as a leader in AI transparency regulation, following its earlier Consumer Genitive AI Transparency Act and establishing mechanisms to assess frontier model safety before deployment.
Source: Cooley Law Firm Analysis →China requires explicit and implicit labeling of all AI-generated content distributed through digital platforms. Major platforms including WeChat, Douyin, Weibo, and others must implement detection systems to identify AI-generated content and apply appropriate labels. Non-compliance results in significant penalties.
The requirement establishes standardized transparency practices for AI-generated content across Chinese digital platforms and reflects ongoing efforts to track and manage AI-generated information online.
Source: CGTN →The first enforcement wave of the EU AI Act becomes law. Social credit scoring, real-time biometric identification in public spaces, and other high-risk AI applications are prohibited. Fines of up to 30 million euros or 6% of annual global revenue apply.
This is the most significant AI enforcement action globally, establishing hard legal restrictions and creating a precedent for other jurisdictions.
Read more →South Korea's legislature approves the AI Basic Act, establishing comprehensive framework for AI development and deployment. The law addresses transparency, safety, and fairness in AI systems while promoting innovation.
South Korea becomes one of the few nations with standalone AI legislation, alongside the EU, signaling Asia's regulatory ambitions.
Read more →Brazil's AI regulation bill, focusing on transparency, accountability, and risk-based governance, advances through Congress. The proposed law includes requirements for disclosure of AI-generated content and liability frameworks for AI systems.
Brazil is positioning itself as a major voice in global AI governance, balancing innovation with consumer protection in Latin America.
Read more →Multiple US states including Colorado, California, and others advance AI-related legislation covering transparency, algorithmic auditing, and consumer protection. A patchwork of state-level regulations emerges.
Without federal AI legislation, the US creates a fragmented regulatory landscape, contrasting with the unified EU approach.
Read more →Australia publishes a voluntary AI safety standard as part of its broader AI regulatory strategy. The framework emphasizes responsible AI use and provides guidance for organizations deploying AI systems.
Australia adopts a lighter regulatory touch compared to the EU, focusing on industry self-regulation and standard-setting rather than hard legal restrictions.
Read more →The EU AI Act becomes enforceable law across all Member States. While some prohibitions apply immediately, a phased implementation period extends through 2025 for other requirements. Compliance obligations begin for high-risk AI systems and providers.
This marks the world's first comprehensive AI legislation, establishing the legal framework that governments worldwide are using as a reference for their own regulations.
Read more →Spain establishes AESIA (Agencia Española de Supervisión de IA), the first dedicated national AI oversight agency in the European Union. Headquartered in A Coruña, AESIA coordinates AI compliance and enforcement across Spain and supports broader EU AI governance efforts.
Spain's creation of AESIA marks a significant institutional development, establishing a model for how member states can build specialized AI regulatory capacity within the EU's federated framework.
Source: Spanish Government →Colorado becomes the first US state to pass comprehensive AI consumer protection legislation with Senate Bill 24-205. The law requires AI developers to prevent algorithmic discrimination and discrimination in decisions affecting consumer rights. Originally scheduled to take effect February 1, 2026, the effective date was extended to June 1, 2026.
Colorado establishes itself as a pioneer in US state-level AI regulation, creating a model that other states including California and New York have followed with their own frontier AI laws.
Source: Colorado Legislature →The European Parliament formally adopts the AI Act following extensive negotiation and compromise. The law establishes a risk-based approach to AI regulation, with prohibited categories, high-risk systems, and transparency requirements.
Adoption signals the finalization of the world's first comprehensive AI legislation and sets the stage for implementation and enforcement.
Read more →The European Commission establishes the AI Office as a dedicated body to implement and enforce the AI Act. The office coordinates across Member States and serves as the primary authority for AI regulation enforcement.
The creation of a dedicated regulatory body signals the EU's commitment to active, ongoing AI governance rather than one-time legislation.
Read more →The UK hosts the first international AI Safety Summit at Bletchley Park, bringing together government leaders, AI researchers, and industry experts from 28 countries. Discussions focus on frontier AI safety and global governance approaches.
The summit affirms AI safety as a global priority and establishes the UK as a key voice in international AI governance discussions.
Read more →The G7 nations release the Hiroshima AI Process Code of Conduct, a voluntary framework for AI governance developed during Japan's presidency. Major AI companies sign on to commitments around transparency and safety.
The code represents international consensus on AI governance principles, though critics note its voluntary nature limits enforceability.
Read more →President Biden issues an Executive Order on AI Safety and Standards, directing federal agencies to establish AI safety standards, conduct audits, and develop best practices. The order focuses on high-impact AI systems and critical infrastructure.
The Executive Order represents the most significant US federal action on AI governance to date, though it lacks the legal force of legislation.
Read more →China's regulations on generative AI services become effective, requiring content security assessments, algorithm transparency, and state security reviews. AI systems must comply with national content standards and cannot generate prohibited content.
China establishes a distinct national framework for AI governance, reflecting its existing regulatory approach to technology and online services.
Read more →After reaching a settlement on data protection concerns, Italy lifts its ban on ChatGPT. OpenAI implements age verification and data processing commitments to comply with Italian and EU data protection standards.
The ban resolution demonstrates how individual nations can pressure major AI providers to improve compliance, even without comprehensive legislation.
Read more →India's government initially signals a preference for AI innovation over heavy regulation, resisting calls for immediate comprehensive legislation. This reflects India's focus on AI as an economic growth driver.
India's approach contrasts with EU and other jurisdictions, emphasizing growth while committing to monitoring and advisory frameworks rather than legal mandates.
Read more →Italy's Data Protection Authority temporarily bans ChatGPT due to concerns about data collection, age verification, and GDPR compliance. OpenAI is given 30 days to comply or face fines of up to 20 million euros or 4% of revenue.
Italy's ban becomes a significant moment, highlighting privacy concerns with AI systems and demonstrating that individual countries can enforce restrictions absent EU-wide AI legislation.
Read more →The National Institute of Standards and Technology (NIST) releases the AI Risk Management Framework, a voluntary guideline for managing AI risks across development, deployment, and monitoring. The framework provides practical standards without legal mandates.
The NIST framework becomes a widely-adopted industry standard and resource, influencing both private sector and regulatory approaches globally.
Read more →China's regulations on deepfake and synthetic media (deep synthesis) become law, requiring content labels and provider registration. The rules target deepfakes, generated audio, and other synthetic media that could spread false information.
China becomes an early mover in regulating specific AI applications, particularly those affecting information integrity and content authenticity.
Read more →Canada's government introduces the Artificial Intelligence and Data Act (AIDA), proposed legislation establishing a risk-based framework for AI governance. The law would require impact assessments and transparency for high-risk AI systems.
Canada positions itself as an early adopter of comprehensive AI legislation, though passage remains pending as of 2025.
Read more →UNESCO's Members adopt the Recommendation on AI Ethics, establishing international norms for responsible AI development and deployment. The recommendation emphasizes human rights, transparency, and accountability in AI systems.
The UNESCO recommendation provides non-binding international guidance that influences national policy discussions and establishes shared ethical principles.
Read more →The European Commission formally proposes the AI Act, the world's first comprehensive AI legislation. The proposal introduces a risk-based framework, prohibited categories, and high-risk AI system requirements, marking the beginning of the legislative process.
The proposal launches what will become the most comprehensive AI regulation globally and signals the EU's commitment to a rights-based approach to AI governance.
Read more →This page documents publicly reported regulatory actions for informational and educational purposes. All descriptions are based on published news reporting and official government communications. This is not legal advice.
Some content on this page was created with the assistance of AI tools.