The Brussels Effect is the mechanism by which EU regulation becomes a de facto global standard even without any treaty or coordinated adoption by other jurisdictions. The term was coined by Anu Bradford in "The Brussels Effect," Northwestern University Law Review 107:1 (2012), developed at book length in The Brussels Effect (2020), and revisited in *Digital Empires* (2023). When the conditions hold, firms find it cheaper to apply EU rules to their products worldwide than to maintain region-specific behavior, so EU rules are exported globally.
Conditions
Bradford identifies five conditions that together produce the effect:
- Market size — the EU is too large to forgo, so firms must comply for EU customers.
- Regulatory capacity — EU regulators have credible enforcement.
- Stringent rules — EU rules are typically stricter than alternative regimes.
- Inelastic targets — the regulated conduct cannot easily be moved out of the EU.
- Non-divisibility — compliance is cheaper when applied uniformly globally than when bifurcated by region. This is the "California Effect" property, in which a firm that changes its product for a large regulated market often changes it everywhere.
When all five hold, EU rules export worldwide whether or not other jurisdictions endorse them.
GDPR as paradigmatic case
The 2018 General Data Protection Regulation (GDPR) is the framework's paradigmatic case. Global platforms including Google, Meta, and Microsoft adopted GDPR-style consent and data-subject-rights flows worldwide rather than maintain region-specific behaviors. US state privacy laws, including California's CCPA and Virginia's VCDPA, borrow GDPR structure.
EU AI Act as the AI-era test case
The EU AI Act (2024 onward) is the current test of whether the effect extends to AI. Its obligations on general-purpose AI (GPAI) providers apply to any model placed on the EU market; for frontier labs, this includes Anthropic, OpenAI, Google DeepMind, xAI, and Meta, all of whom the AI Office treats as GPAI providers. The AI Office enforcement and GPAI guidelines describe the operational enforcement posture.
Evidence of a partial Brussels Effect in AI to date includes several developments. The GPAI Code of Practice has voluntary signatories that include US frontier labs, enabling a presumption of compliance. Safety-case documentation submitted under the AI Act's safety and security framework draws on materials structurally similar to what US labs already produce, such as their Responsible Scaling Policies and Preparedness Frameworks. US state laws including California SB 53 and the New York RAISE Act sometimes echo AI Act structure.
Several factors limit the effect in AI specifically. Divisibility is partially present, because labs can deploy region-specific model variants more cheaply than firms can produce region-specific physical goods. US federal preemption, via Trump-era AI export executive orders, actively pushes against EU regulatory reach on US firms. Chinese AI rules, including the generative-AI interim measures among others, create a third regulatory pole rather than an EU–US duopoly.
Relation to the three-empires framework
Bradford's three-empires framework treats the Brussels Effect as the export mechanism of the European regulatory model, contrasting it with the US market-driven export of technology and products and China's state-driven export of digital infrastructure. The AI-era question is whether EU regulation or US industrial policy, as in America's AI Action Plan and OpenAI's industrial-policy positioning, shapes the global diffusion frontier.
Relationships
- depends-on: Digital Empires — Anu Bradford (2023) — the book-length treatment.
- supports: EU AI Act (Regulation 2024/1689), EU AI Office — Enforcement Framework + GPAI Provider Guidelines, EU General-Purpose AI Code of Practice (Final Version, 2025) — the current AI-specific test cases.
- supports: Techno-Federalism: How Regulatory Fragmentation Shapes the U.S.-China AI Race — Wu's framework builds on Bradford.
- contradicts: America's AI Action Plan / Executive Order 14320 — Promoting the Export of the American AI Technology Stack — US regulatory counter-posture explicitly rejects EU-style rules as export vector.
- related: The Digitalist Papers (Stanford, Volumes 1–2) — Cochrane and Volokh essays engage critically with the EU regulatory model.
- related: Incentives or Obligations? The U.S. Regulatory Approach to Voluntary AI Governance Standards — US voluntary-standards approach as the alternative to Brussels Effect mandatory-regulation export.