Over the last decade, it has become fashionable to pour scorn on the EU’s power to shape global tech markets. The United States builds the platforms. China scales the infrastructure. What right does Europe have, its critics say, to back-seat drive — imposing rules for technologies in which it is not even commercially competitive?
In its heyday, EU regulations like the GDPR set the global standard. If a company touched the data of people in the EU, the law reached them no matter where they were based. The penalties were big enough to command real boardroom attention.
More often than not, pragmatism set in and multinational firms deemed it more efficient to adopt GDPR-style standards across their operations rather than manage different privacy rules in different countries.
In many ways, this was the apex of the so-called ‘Brussels Effect’ — the term coined for EU regulation’s uncanny ability to become the de facto global rulebook.
But as the gap widened between Europe and its rivals in the commercial tech race, political backlash was inevitable. In Washington, Brussels came to be cast less as a global standard-setter than as a hostile regulator targeting American firms.
The backlash against Brussels
Regulations like the Digital Services Act in 2022, designed to govern platform content moderation, were viewed by many across the pond as an affront to the US First Amendment right to free speech.
As a result, US firms were increasingly motivated to resist. Some challenged Brussels directly in court; others redesigned only their EU-facing services or carved out Europe-specific product versions, instead of exporting the rules across their entire business.
As we hurtle towards the next deadline for companies to comply with the EU AI Act this August, however, there is growing evidence to suggest it may be too soon to write off the Brussels Effect just yet.
A new analysis of the world’s largest dataset on corporate AI use from the Thomson Reuters Foundation — sampling almost 3,000 global companies, shows that the AI Act is already generating a global ripple effect.
Nearly half of the companies that cite the EU AI Act in their disclosures are headquartered outside the EU. Even more strikingly, the United States is the single largest source of non-EU companies engaging with the regulation.
So, what might have changed? On the face of it, not much. AI is probably the clearest illustration of Europe’s commercial lag behind the US. And the political opposition is also still very much there. The White House has framed AI regulation as a threat to US technological leadership.
Why the AI Act may travel further
But the AI Act does have several tricks up its sleeve that may explain why it could be on course to become the global default.
First, it is tied to market access. If a company wants an AI-powered product that can be sold or deployed across the entirety of the EU single market, it must comply with the rules.
It is also extraterritorial. The Act applies not just to firms established in the EU, but also to providers placing an AI product on the EU market from a third country, and even to providers in a third country where the AI product’s output is used in the Union. That makes ‘ignore Europe’ a difficult strategy for globally active firms.
And then there’s its teeth. The AI Act carries far more punitive penalties than most other EU tech regulations, with the most serious breaches attracting fines of up to €35m or 7 percent of total worldwide annual turnover.
For global companies, the prospect of running separate AI governance systems for different markets may well prove too operationally complex, legally risky and commercially inefficient. Given the Act’s comparatively high level of stringency relative to that of other markets, it is simply safer to adopt it as their global benchmark.
The real test is accountability
Yet the debate around the Act rages on and remains trapped between two false extremes. On one side are those who argue that regulation will inevitably suffocate innovation. On the other are those who imply that regulation alone can make AI trustworthy.
The absence of strong rules is not a pro-innovation position. Trustworthy AI will not emerge from market forces alone, particularly when harms are often opaque and borne by those with the least power to contest them.
At the same time, poorly designed regulation can certainly create burdens, especially for smaller companies and civil society organisations that lack the compliance teams of larger players.
The real test for the AI Act will not be whether companies mention it in their policies. It will be whether affected people can understand when AI is being used, challenge harmful outcomes, and rely on meaningful accountability when things go wrong.
Influence is not the same as impact. If Europe wants its AI rulebook to become a global model, it must prove that the rules do more than travel. They must protect people.
