Brussels has long been accused of overregulation, but new data suggests its approach to artificial intelligence is having an unintended consequence: shaping corporate behavior far beyond the bloc's borders. Nearly half of all companies that reference the EU's Artificial Intelligence Act in their governance disclosures are not legally required to comply with it, according to research published by the Thomson Reuters Foundation.
The analysis, drawn from more than 100,000 data points across 2,973 companies worldwide, found that 47% of firms citing the Act are headquartered outside the EU. This pattern echoes what scholars call the "Brussels Effect" — the tendency for EU regulation to become a de facto global benchmark, as happened with the General Data Protection Regulation (GDPR).
A phased rollout with extraterritorial teeth
The AI Act, which entered into force in 2024, is the world's first comprehensive cross-sector AI law. Its obligations are being phased in gradually: bans on the highest-risk AI uses and transparency rules for general-purpose models — including requirements to label deepfakes and chatbots — took effect in December 2025. The most consequential provisions, covering high-risk systems used in hiring, credit, and healthcare, become fully binding in August 2026.
Like the GDPR, the Act applies extraterritorially. Any organization whose AI systems are used within the EU or whose outputs affect EU citizens, businesses, or public institutions must comply, regardless of where it is based. Penalties for the most serious breaches can reach €35 million or 7% of global annual revenue.
Yet the research also reveals a broader gap in corporate AI governance. Across all sectors, only 13% of companies have any formal AI governance framework at all. Of that minority, just over half (53%) specifically reference the EU AI Act. And within that group, 47% are non-EU firms.
Tech and finance lead the charge
Engagement with the Act varies sharply by industry and geography. Information technology companies account for nearly 40% of all non-EU firms citing the legislation, with communication services and financial services adding another 29%. Regionally, North America leads at just under 40%, driven largely by US technology and healthcare companies with significant EU market presence. Non-EU European firms — particularly from the United Kingdom, Switzerland, and Norway — follow at around 24%, reflecting close commercial and regulatory ties to the bloc. Asian companies, concentrated in global AI supply chains, account for roughly 28%.
The United States offers a particularly striking case. Despite having no overarching federal AI law, American companies represent 35% of all non-EU citers — the single largest national contributor. Within the US, 53% of citing firms come from the IT sector, and one in five US IT companies in the dataset references the Act, the highest rate of any sector nationally. Major US technology firms including Microsoft, Google, OpenAI, and xAI have voluntarily aligned with elements of the EU's AI Code of Practice, motivated by continued access to the European market.
Strengths and blind spots
Companies that mention the Act tend to perform well on basic governance metrics: they have clear AI strategies, board-level oversight, and transparency about data use. Non-EU firms citing the Act even outperform EU companies on these measures. However, EU firms lead on workforce training, with 49.4% offering reskilling or AI literacy programs, compared with 40.6% of non-EU firms.
But strategic oversight does not always translate into operational safeguards. Only 12.4% of companies worldwide have a policy requiring human review of individual AI decisions, and nearly half of those have not yet implemented it in practice. Rights checks are even rarer: fewer than one in four companies assess whether their AI could harm employee rights, even among those most engaged with the Act.
The findings underscore that while the EU AI Act is reshaping corporate rules from Washington to Tokyo, its influence remains concentrated in sectors and regions with strong market incentives to align. As the August 2026 deadline approaches, the gap between aspiration and implementation may narrow — but for now, the Brussels Effect is real, significant, and far from universal.
