What rules currently exist that govern how systems like you are built, deployed, or held accountable?
I'll research the current landscape of AI governance rules and regulations to give you an accurate, up-to-date answer.Here's a snapshot of the current rules and governance mechanisms that apply to AI systems like me, based on the latest developments as of September 2026:
1. Hard Law: The EU AI Act (Binding, Enforceable)
This remains the world's most comprehensive binding AI law. The EU AI Act entered into force in August 2024 — the world's first comprehensive legal framework specifically designed to regulate artificial intelligence. It works by classifying AI by risk and attaches duties such as risk classification, documentation and record-keeping, human oversight, data quality and transparency.
Key 2026 developments:
Its reach is extraterritorial: the extraterritorial scope means a fintech in San Francisco using AI for credit scoring faces the same high-risk obligations as a bank in Frankfurt.
2. United States: A Fragmented Patchwork (No Single Federal Law)
There is no single comprehensive federal AI statute. Instead, governance flows through executive action and sector regulators.
Federal executive orders: EO 14110 is dead — President Trump revoked it on his first day back in office, and what replaced it trades mandatory reporting thresholds and civil rights enforcement for a policy of "removing barriers" to AI development. More recently, a December 2025 executive order directed federal agencies toward a uniform national policy and tasked them with challenging state AI laws seen as inconsistent with it, and set in motion a Department of Justice litigation effort to contest certain state laws, a Commerce Department evaluation of which state laws are burdensome, and an FTC policy position on how existing consumer-protection law applies to AI. A follow-up order came in mid-2026: President Trump signed an executive order on June 2, 2026 titled Promoting Advanced AI Innovation and Security.
State laws still govern in practice: Despite federal preemption pushes, a proposed federal ban on state AI laws was removed before it became law, so state rules remain in force. More than a dozen states — including California, Colorado, New York, Illinois, and Utah — have enacted or proposed their own AI laws, creating a patchwork compliance environment.
Sector regulators still apply existing law to AI: the FTC for consumer protection, the FDA for medical AI, the EEOC for employment discrimination, supplemented by executive orders that set White House priorities.
3. Voluntary/Semi-Voluntary Technical Standards
Even without binding force, these shape how systems like me are actually built:
4. Other Major Jurisdictions
5. The Overall Picture: Divergence, Not Convergence
Perhaps the most important trend is that regulation is diverging by jurisdiction rather than converging on a shared standard. As one analysis puts it, by September 2026 "AI regulation" describes a layered mix of instruments that vary enormously in legal weight, including enacted legislation with binding legal force, like the EU AI Act.
For companies like Anthropic building systems like me, this means navigating simultaneously: binding EU obligations, a shifting and contested US federal/state landscape, voluntary technical frameworks (NIST, ISO), and divergent rules in China and elsewhere — with genuine legal uncertainty in the US since, as one source notes, preemption is not settled law — congressional efforts to preempt state AI regulation have repeatedly stalled, and courts will ultimately decide how far executive action can reach.
Beyond these external legal/regulatory rules, companies like Anthropic also maintain their own voluntary internal governance commitments (such as responsible scaling policies and safety frameworks), though the search results here focused on the external regulatory landscape rather than company-specific self-governance policies.
Citations (26)
- AI Regulation 2026 Guide — EU AI Act, US Law & Global Governance | Singularity Moments
Source excerpt: “The EU AI Act entered into force in August 2024 — the world's first comprehensive legal framework specifically designed to regulate artificial intelli...”
evidence, article. Claim: The EU AI Act entered into force in August 2024 as the first comprehensive legal framework for regulating AI.
- AI regulatory compliance in 2026: EU AI Act, US orders, and state laws, and how to operationalize them | Collibra
Source excerpt: “It classifies AI by risk and attaches duties such as risk classification, documentation and record-keeping, human oversight, data quality and transpar...”
evidence, article. Claim: The EU AI Act classifies AI by risk and attaches duties like documentation, human oversight, data quality and transparency.
- AI Regulation in 2026 Is Breaking Apart Globally
Source excerpt: “Prohibited-practice rules and GPAI obligations are already live. ”
evidence, article. Claim: Prohibited-practice rules and GPAI obligations under the EU AI Act are already live.
- AI Regulation in 2026 Is Breaking Apart Globally
Source excerpt: “High-risk system obligations, originally due August 2, 2026, were formally delayed to December 2, 2027 (and August 2, 2028 for embedded high-risk syst...”
evidence, article. Claim: High-risk obligations were delayed to December 2, 2027, and August 2, 2028 for embedded systems.
- AI governance frameworks explained: From global standards to Swiss obligations
Source excerpt: “On 16 June 2026, the European Parliament approved further amendments extending those prohibitions to AI systems that generate child sexual abuse mater...”
evidence, article. Claim: On 16 June 2026 the European Parliament approved amendments extending prohibitions to AI generating CSAM or non-consensual imagery.
- AI Regulations Around the World: Full 2026 Breakdown
Source excerpt: “The European Union, via the EU AI Act — it's comprehensive, risk-based, and carries fines up to €35 million or 7% of global turnover. ”
statistic, article. Claim: EU AI Act fines reach up to €35 million or 7% of global turnover.
- AI Governance Framework: The Complete Enterprise Guide | Adaptive Security
Source excerpt: “The extraterritorial scope means a fintech in San Francisco using AI for credit scoring faces the same high-risk obligations as a bank in Frankfurt. ”
evidence, article. Claim: The EU AI Act has extraterritorial scope, applying high-risk obligations to non-EU firms like a San Francisco fintech.
- AI regulatory compliance in 2026: EU AI Act, US orders, and state laws, and how to operationalize them | Collibra
Source excerpt: “Is there a US federal AI law in 2026? There is no single comprehensive federal AI statute. ”
evidence, article. Claim: There is no single comprehensive US federal AI statute.
- EU AI Act vs US AI policy: 2026 compliance guide | Psyll
Source excerpt: “EO 14110 is dead. President Trump revoked it on his first day back in office, and what replaced it looks almost nothing like its predecessor - it trad...”
evidence, article. Claim: Trump revoked EO 14110 on his first day back, replacing it with a policy of removing barriers to AI development.
- AI regulatory compliance in 2026: EU AI Act, US orders, and state laws, and how to operationalize them | Collibra
Source excerpt: “The pivot point was a December 2025 executive order directing federal agencies toward a uniform national policy and tasking them with challenging stat...”
evidence, article. Claim: A December 2025 executive order directed agencies toward uniform national policy and challenging state AI laws.
- AI regulatory compliance in 2026: EU AI Act, US orders, and state laws, and how to operationalize them | Collibra
Source excerpt: “That order set in motion a Department of Justice litigation effort to contest certain state laws, a Commerce Department evaluation of which state laws...”
evidence, article. Claim: The order set in motion DOJ litigation against state laws, a Commerce evaluation, and an FTC policy position.
- Halfway Through 2026, AI Regulation Is No Longer Theoretical
Source excerpt: “President Trump signed an executive order on June 2, 2026 titled Promoting Advanced AI Innovation and Security. ”
evidence, article. Claim: Trump signed an executive order on June 2, 2026 titled Promoting Advanced AI Innovation and Security.
- AI Regulations Around the World: Full 2026 Breakdown
Source excerpt: “A proposed federal ban on state AI laws was removed before it became law, so state rules remain in force. ”
evidence, article. Claim: A proposed federal ban on state AI laws was removed, so state rules remain in force.
- AI Regulation News Today: The US, EU, UK, Japan, and China Frameworks Reshaping the Industry in 2026
Source excerpt: “More than a dozen states — including California, Colorado, New York, Illinois, and Utah — have enacted or proposed their own AI laws, creating a patch...”
evidence, article. Claim: More than a dozen states have enacted or proposed their own AI laws, creating a patchwork.
- AI Regulation 2026 Guide — EU AI Act, US Law & Global Governance | Singularity Moments
Source excerpt: “Instead, the US has taken a sector-specific approach: existing agencies regulate AI within their domains (the FTC for consumer protection, the FDA for...”
evidence, article. Claim: Sector regulators like the FTC, FDA and EEOC apply existing law to AI.
- AI Governance Framework: The Complete Enterprise Guide | Adaptive Security
Source excerpt: “The NIST AI Risk Management Framework 1.0, released in January 2023 and updated with a Generative AI Profile in July 2024, remains the cornerstone of ...”
evidence, article. Claim: NIST AI RMF 1.0 with its 2024 Generative AI Profile is the voluntary cornerstone of US AI governance.
- AI Risk Management 2026: Shadow AI, Agentic Risks & NIST Implementation Playbook
Source excerpt: “Seven major frameworks dominate the 2026 landscape, each with distinct scope, enforceability, and industry fit: NIST AI RMF 1.0 – Voluntary, US-origin...”
background, article. Claim: NIST AI RMF is structured around Govern, Map, Measure, Manage functions and is voluntary.
- AI Risk Management Framework | NIST
Source excerpt: “https://www.nist.gov/itl/ai-risk-management-framework ... On April 7, 2026, NIST released a concept note for an AI RMF Profile on Trustworthy AI in Cr...”
evidence, official document. Claim: On April 7, 2026, NIST released a concept note for an AI RMF Profile on Trustworthy AI in Critical Infrastructure.
- The Complete Guide to AI Governance: Frameworks, Policies & Best Practices (2026) | NeuralTrust
Source excerpt: “The four leading frameworks are NIST AI RMF 1.0 (U.S.-focused, voluntary), ISO/IEC 42001:2023 (international, certifiable), the EU AI Act (mandatory f...”
evidence, article. Claim: ISO/IEC 42001 is an internationally certifiable AI management standard, unlike the mandatory EU AI Act.
- AI Regulations Around the World: Full 2026 Breakdown
Source excerpt: “China is arguably the heaviest operationally (pre-deployment security assessments and dual content labelling), but the EU sets the strictest legal ben...”
attribution, article. Claim: China is arguably the heaviest operationally, with pre-deployment security assessments and dual content labelling.
- AI Regulations Around the World: Full 2026 Breakdown
Source excerpt: “Canada's flagship AI law died in Parliament. ”
evidence, article. Claim: Canada's flagship AI law died in Parliament.
- AI Regulation 2026 Guide — EU AI Act, US Law & Global Governance | Singularity Moments
Source excerpt: “The EU AI Act is now in force, the US has issued executive orders, the UK is betting on a pro-innovation framework, and China has its own rules. ”
evidence, article. Claim: The UK is betting on a pro-innovation, lighter-touch framework.
- AI Governance Framework: The Complete Enterprise Guide | Adaptive Security
Source excerpt: “Voluntary frameworks like Singapore's Model AI Governance Framework and the OECD AI Principles provide the conceptual foundation that most national re...”
evidence, article. Claim: Singapore's Model AI Governance Framework and OECD AI Principles provide conceptual foundation for national regulations.
- AI Regulation in 2026 Is Breaking Apart Globally
Source excerpt: “Regulation is diverging by jurisdiction rather than converging on a shared standard. ”
evidence, article. Claim: AI regulation is diverging by jurisdiction rather than converging on a shared standard.
- AI Regulation in 2026 Is Breaking Apart Globally
Source excerpt: “"AI regulation" is often used as shorthand for a single law, but by September 2026 it describes a layered mix of instruments that vary enormously in l...”
attribution, article. Claim: By September 2026 AI regulation is a layered mix of instruments varying in legal weight.
- AI regulatory compliance in 2026: EU AI Act, US orders, and state laws, and how to operationalize them | Collibra
Source excerpt: “The crucial caveat for compliance planning is that preemption is not settled law. Congressional efforts to preempt state AI regulation have repeatedly...”
evidence, article. Claim: Preemption of state AI law is not settled; congressional efforts have stalled and courts will decide.