Direct comparison
EU AI Act vs US AI regulation
The EU AI Act and US AI regulation represent two contrasting models: one comprehensive, binding and risk-based; the other a sectoral and state-level patchwork with voluntary federal frameworks.
The step most authors miss
Doing CRediT right? Don’t stop at the statement.
A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.
Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.
Side-by-side comparison
| Dimension | EU AI Act | US AI regulation |
|---|---|---|
| What it is | A single comprehensive, horizontal law — Regulation (EU) 2024/1689 — governing AI across all sectors of the Union. | A patchwork of sectoral rules, state and city laws, agency guidance and voluntary federal frameworks; no comprehensive federal AI statute. |
| Legal force | Binding and directly applicable in all EU Member States without national transposition. | Mixed: binding state and city laws apply locally, while leading federal frameworks (NIST AI RMF) are voluntary. |
| Core approach | Risk-based tiers: prohibited, high-risk, limited-risk and minimal-risk, with obligations scaling to risk. | Largely use- and sector-specific, focused on particular harms such as discrimination in hiring or lending. |
| Scope | Horizontal — applies across industries to providers and deployers, with extraterritorial reach to non-EU actors serving the EU market. | Fragmented by jurisdiction and sector; coverage and definitions differ between states and cities. |
| Prohibited uses | Explicit Article 5 bans, e.g. social scoring by public authorities and untargeted facial-image scraping. | No single national ban list; some states (e.g. Texas TRAIGA) prohibit specific uses such as intentional unlawful discrimination. |
| High-risk regime | Defined Annex III domains trigger strict duties: risk management, data governance, human oversight, conformity assessment, CE marking. | No uniform high-risk regime; Colorado’s AI Act applies a reasonable-care duty to high-risk systems in consequential decisions. |
| General-purpose AI | Dedicated GPAI obligations, with extra duties for models posing systemic risk, supported by a Code of Practice. | No dedicated federal GPAI statute; addressed through voluntary commitments, guidance and sectoral measures. |
| Federal vs sub-national lead | EU-level legislation provides one harmonised framework across Member States. | States and cities lead with binding rules; the federal level emphasises voluntary frameworks and agency action. |
| Penalties | Tiered administrative fines up to €35M or 7% of global annual turnover for prohibited-practice breaches. | Varies by law and jurisdiction; enforced through state attorneys general, agencies or existing statutes rather than one penalty scheme. |
Common questions
FAQ
Is US AI regulation equivalent to the EU AI Act?+
No. The EU AI Act is a single comprehensive, binding, risk-based regulation applying across the Union. US AI regulation is a patchwork of sectoral rules, state and city laws and voluntary federal frameworks such as the NIST AI RMF, without an equivalent comprehensive federal statute.
Does the EU AI Act apply to US companies?+
It can. The EU AI Act has extraterritorial reach: it may apply to providers and deployers based outside the EU, including US companies, where their AI systems are placed on the EU market or their output is used within the Union.
What is the NIST AI RMF in the US approach?+
The NIST AI Risk Management Framework is a voluntary US federal framework that helps organisations manage AI risks. It is not a binding law, but it functions as a common reference point across sectors and states, in contrast with the EU AI Act’s binding requirements.
Going deeper
