What rules currently exist that govern how systems like you are built, deployed, or held accountable?

AI governance is a fast-moving area, so let me get current information on the key regulatory frameworks.The rules governing AI systems like me fall into several overlapping categories: binding law, executive/administrative action, company-level policy, and voluntary international frameworks. Here's the current landscape as of September 2026.

1. European Union: The AI Act (the most comprehensive binding framework)

The EU AI Act, adopted on 21 May 2024, lays down harmonized rules on artificial intelligence and is the first-ever comprehensive legal framework on AI worldwide. Its approach is risk-based: the EU's regulatory model is founded on a graduated, risk-oriented structure, differentiating AI systems by potential harm rather than targeting specific technologies, with escalating obligations as risk increases. AI systems that significantly impact fundamental rights are either prohibited or subject to stricter requirements and human oversight.

Key obligations now in force or phasing in:

•As of August 2, 2026, the Act's most consequential obligations took effect: Annex III high-risk AI system requirements, Article 50 transparency obligations, conformity assessments, CE marking, and AI Office enforcement powers.

•However, a 2026 "Digital Omnibus" reform pushed back some deadlines: high-risk obligations for stand-alone systems covering hiring, credit scoring, education, and critical infrastructure are deferred to 2 December 2027, while high-risk AI embedded in already-regulated products like medical devices is deferred to 2 August 2028.

•Transparency rules were not delayed: the transparency duties (Article 50) and the AI Office's enforcement powers over general-purpose AI providers took effect exactly as scheduled on 2 August 2026.

•For general-purpose AI models like me, providers must maintain detailed technical documentation covering model architecture and training procedures, provide that documentation to downstream integrators, publish a summary of training data content, and comply with EU copyright law including text-and-data-mining opt-outs.

•Transparency requirements mean chatbot systems must clearly disclose their AI nature before or during interaction, unless it's already obvious to a reasonable user.

2. United States: A shifting, deregulation-leaning federal posture + growing state patchwork

Federal level: The Trump administration has moved away from binding federal AI safety rules toward a lighter-touch, innovation-first approach:

•On his first day in office, Trump revoked Biden's 2023 AI safety executive order (EO 14110).

•In December 2025, Trump signed Executive Order 14365, "Ensuring a National Policy Framework for Artificial Intelligence," aimed at sustaining US AI dominance through a minimally burdensome national policy framework and curbing state-level AI regulation. This order created a DOJ task force and used federal funding leverage, though the final text expressly prohibits federal preemption of state laws on topics like child safety, AI compute infrastructure, and state government use of AI.

•In June 2026, a follow-up order titled "Promoting Advanced Artificial Intelligence Innovation and Security" established new cybersecurity mandates and a voluntary framework for the secure deployment of frontier AI models. It also directs the Attorney General to prioritize enforcement of the Computer Fraud and Abuse Act against those who use AI to unlawfully access computer systems.

•Crucially, since Congress has not passed a federal law preempting state AI laws, existing state laws remain in effect in the short term.

State level: A rapidly growing patchwork now covers accountability and disclosure:

•Colorado's AI Act mandates "reasonable care" to prevent algorithmic discrimination in high-risk AI systems, with major provisions effective mid-2026.

•Colorado also passed a chatbot-specific law: the Chatbot Safety Act requires operators to estimate user age, disclose that users are interacting with AI rather than humans, safeguard teen users against explicit content and simulated emotional dependence, and provide privacy/account tools for minors, plus requires suicide/self-harm response protocols and prohibits representing chatbot outputs as equivalent to licensed professional services.

•California's framework is one of the most developed: the California AI Transparency Act (SB 942, as amended by AB 853) requires covered generative-AI providers with over 1 million monthly California users to embed C2PA-compatible provenance watermarks in AI-generated images, video, and audio, and to offer a free public detection tool. Separately, the Generative AI Training Data Transparency Act (AB 2013) requires developers of publicly available generative AI systems to publish a summary of their training data.

•Companion-chatbot laws have spread fast: 2026 started with three states regulating companion chatbots; by mid-2026 the count reached 12, with new chatbot laws enacted in Colorado, Connecticut, Georgia, Hawaii, Idaho, Iowa, Nebraska, Oregon, and Washington.

3. China: Binding, sector-specific rules built around content and security

China has no single comprehensive AI statute but instead layers targeted binding rules:

•The cornerstone statutes are the Cybersecurity Law, Data Security Law, and Personal Information Protection Law, layered with AI-specific rules: the Algorithm Recommendation Measures (2022), Deep Synthesis Measures (2023), and Interim Measures for Generative AI Services (2023) — making China the first country to implement binding generative AI regulations.

•Service providers are held responsible for content generated using their services.

•More recently, Measures for Labelling AI-Generated or Synthesised Content require providers to attach explicit labels to AI-generated text, images, audio, and video, and embed implicit labels in file metadata.

•China has also issued non-binding guidance: China's AI Safety Governance Framework 2.0, published in 2025, provides structured guidance on risk categorisation and countermeasures across the AI development and deployment process.

4. Voluntary international/industry frameworks

Beyond hard law, a set of non-binding but influential frameworks shape how frontier AI developers (including companies like Anthropic) operate:

•Several regional and interregional governance frameworks now articulate shared expectations for managing risks from general-purpose AI, providing non-binding guidance for policymakers and organisations.

•Since early 2025, the risk-management landscape has evolved with new resources including the EU's General-Purpose AI Code of Practice, the G7 Hiroshima AI Process (HAIP) Reporting Framework, China's AI Safety Governance Framework 2.0, and various AI developers' own Frontier AI Safety Frameworks.

•Expert-led initiatives such as the Singapore Consensus, developed by AI scientists from multiple countries, outline research priorities for general-purpose AI safety across risk assessment, development, and control.

•Individual companies also publish their own governing principles — for instance, OpenAI has guidelines focusing on long-term safety, robustness, and compliance with ethical standards. Anthropic similarly operates under its own published Responsible Scaling Policy and usage policies, which function as a form of voluntary self-governance layered on top of these external legal regimes.

Bottom line

No single law governs a model like me end-to-end. Instead, accountability comes from a patchwork: binding EU rules (risk classification, transparency, documentation, and GPAI-specific duties), a US federal posture currently favoring deregulation paired with an expanding thicket of US state laws (especially on chatbot disclosure, minor safety, and algorithmic discrimination), binding Chinese content/security rules, and a growing set of voluntary international and company-level frameworks (G7 HAIP, Frontier Safety Frameworks, etc.) that fill gaps where hard law hasn't yet caught up. This landscape is moving quickly and unevenly across jurisdictions, so specifics are likely to keep shifting.

Citations (29)

1legalnodes.comCompanies
EU AI Act 2026 Updates: Compliance Requirements and Business Risks
Source excerpt: “Adopted on 21 May 2024, the EU Artificial Intelligence Act lays down harmonized rules on artificial intelligence and is the first-ever comprehensive l...”
2legalnodes.comCompanies
EU AI Act 2026 Updates: Compliance Requirements and Business Risks
Source excerpt: “The EU’s regulatory model for artificial intelligence is founded on a graduated, risk-oriented structure. Rather than targeting specific technologies,...”
3legalnodes.comCompanies
EU AI Act 2026 Updates: Compliance Requirements and Business Risks
Source excerpt: “AI systems which significantly impact fundamental rights, are therefore either prohibited or subject to stricter requirements and human oversight. ”
4responsibleailabs.aiCompanies
EU AI Act August 2026: your compliance countdown | RAIL
Source excerpt: “On August 2, 2026, the EU AI Act's most consequential obligations take effect: Annex III high-risk AI system requirements, Article 50 transparenc...”
5softwareimprovementgroup.comCompanies
A comprehensive EU AI Act Summary [August 2026 update] - SIG
Source excerpt: “The Omnibus confirms two fixed dates for the Act’s toughest tier: high-risk obligations for stand-alone systems (Annex III, covering hiring, credit sc...”
6softwareimprovementgroup.comCompanies
A comprehensive EU AI Act Summary [August 2026 update] - SIG
Source excerpt: “Meanwhile, the transparency duties (Article 50) and the AI Office’s enforcement powers over general-purpose AI (GPAI) providers took effect exactly as...”
7decodethefuture.orgBlogs and newsletters
EU AI Act 2026: Penalties, Risk Tiers & New Deadlines
Source excerpt: “Every provider of a GPAI model placed on the EU market must: maintain detailed technical documentation covering model architecture, training procedure...”
8decodethefuture.orgBlogs and newsletters
EU AI Act 2026: Penalties, Risk Tiers & New Deadlines
Source excerpt: “For developers building chatbots or NLP-powered interfaces: this means your system must clearly disclose its AI nature before or during interaction, u...”
9lathropgpm.comCompanies
New Executive Order Signals Evolving Federal Approach to AI - Lathrop GPM
Source excerpt: “On the first day of his second term in office, President Trump revoked President Biden’s 2023 AI safety order (EO 14110). ”
10whitecase.comCompanies
State AI laws under federal scrutiny: Key takeaways from the executive order establishing federal AI policy framework | White & Case LLP
Source excerpt: “On December 11, 2025, President Trump signed an executive order titled “Ensuring a National Policy Framework for Artificial Intelligence,” (“Executive...”
11paulhastings.comCompanies
President Trump Signs Executive Order Challenging State AI Laws | Paul Hastings LLP
Source excerpt: “The final text is also narrowed as it expressly prohibits the federal preemption of otherwise lawful state AI laws, including those relating to child ...”
12lathropgpm.comCompanies
New Executive Order Signals Evolving Federal Approach to AI - Lathrop GPM
Source excerpt: “On June 2, 2026, in a rapidly evolving federal AI policy landscape, President Trump signed a new Executive Order (EO) titled Promoting Advanced Artifi...”
13lathropgpm.comCompanies
New Executive Order Signals Evolving Federal Approach to AI - Lathrop GPM
Source excerpt: “This EO establishes new cybersecurity mandates and a voluntary framework for the secure deployment of frontier AI models. ”
14lathropgpm.comCompanies
New Executive Order Signals Evolving Federal Approach to AI - Lathrop GPM
Source excerpt: “The Attorney General is directed to prioritize enforcement of the Computer Fraud and Abuse Act and other federal criminal statutes against individuals...”
15whitecase.comCompanies
State AI laws under federal scrutiny: Key takeaways from the executive order establishing federal AI policy framework | White & Case LLP
Source excerpt: “While the interaction between federal initiatives and existing state requirements remains uncertain and may vary depending on legislative, regulatory ...”
16orrick.comCompanies
2026 State Chatbot Laws: Key Provisions and Regulatory Trends
Source excerpt: “Colorado AI Act (SB 24-205): Mandates “reasonable care” to prevent algorithmic discrimination in high-risk AI systems; major provisions effective June...”
17coag.govGovernment
Colorado Anti-Discrimination in AI Law (ADAI) Rulemaking | Colorado Attorney General
Source excerpt: “The Chatbot Safety Act amends section 17 to add protections for users of conversational AI services, including requirements for chatbot operators to e...”
18coag.govGovernment
Colorado Anti-Discrimination in AI Law (ADAI) Rulemaking | Colorado Attorney General
Source excerpt: “The Chatbot Safety Act also requires chatbot operators to create suicide and self-harm response protocols and prohibits chatbot outputs from being rep...”
19ailawsbystate.comData and tools
AI Transparency Laws by State: 2026 Requirements
Source excerpt: “California's California AI Transparency Act (SB 942), as amended by AB 853, goes further than training-data disclosure: it requires covered GenAI prov...”
20ailawsbystate.comData and tools
AI Transparency Laws by State: 2026 Requirements
Source excerpt: “California's Generative AI Training Data Transparency Act (AB 2013) requires developers of publicly available generative AI systems to publish a high-...”
21privacyworld.blogCompanies
U.S. AI Law - 2026 Midyear State Update | Privacy World
Source excerpt: “2026 started with three states (California, New Hampshire, and New York) regulating so-called companion chatbots: Now, the count is up to 12. Chatbot ...”
22arxiv.orgAcademic
Who Governs the Machine? A Machine Identity Governance Taxonomy (MIGT) for AI Systems Operating Across Enterprise and Geopolitical Boundaries
Source excerpt: “The three cornerstone statutes are the Cybersecurity Law (CSL, originally 2017), the Data Security Law (2021), and the Personal Information Protection...”
23arxiv.orgAcademic
Who Owns the Knowledge? Copyright, GenAI, and the Future of Academic Publishing
Source excerpt: “Service providers are held responsible for the content generated using their services. ”
24gaicc.orgNot yet classified
China AI Governance Framework: What Global Businesses Need to Know in 2026
Source excerpt: “The Measures for Labelling AI Generated or Synthesised Content require providers to attach explicit labels to AI-generated text, images, audio, and vi...”
25arxiv.orgAcademic
International AI Safety Report 2026
Source excerpt: “China's Al Safety Governance Framework 2.0, published in 2025, provides structured guidance on risk categorisation and countermeasures across the AI d...”
26arxiv.orgAcademic
International AI Safety Report 2026
Source excerpt: “### Broader governance initiatives offer voluntary guidance Several regional and interregional governance frameworks now articulate shared expectatio...”
27arxiv.orgAcademic
International AI Safety Report 2026
Source excerpt: “### **Updates** Since the publication of the last Report (January 2025), the risk management landscape for general-purpose AI has evolved, with the...”
28arxiv.orgAcademic
International AI Safety Report 2026
Source excerpt: “In addition, expert-led initiatives such as the Singapore Consensus, developed by Al scientists from multiple countries, outline research priorities f...”
29arxiv.orgAcademic
Ethical Concerns of Generative AI and Mitigation Strategies: A Systematic Mapping Study
Source excerpt: “OpenAI has similar guidelines focusing on long-term safety, robustness, and compliance with ethical standards [39]. ”