Your data, a battleground: what if US pressure were to play into Europe’s hands?

The watchword came from Brendan Carr, appointed by Donald Trump to head the FCC (Federal Communications Commission, responsible for combating monopolistic practices): “Delete, Delete, Delete”, which is nothing less than a massive deregulation strategy. This phrase sums up quite well the ambition of the US administration since its return to the White House: for every new regulation adopted in the United States, ten must be removed. But this drive for deregulation does not stop at the US borders. It explicitly targets Europe, notably the DSA (Digital Services Act), the DMA (Digital Markets Act), the AI Act (Artificial Intelligence Act) and the GDPR (General Data Protection Regulation). And this is precisely where the matter becomes serious for anyone who handles confidential data in a professional capacity.

A study by the Ifri, published in April 2026 by Mathilde Velliet, documents this offensive in detail. What it reveals goes beyond a mere trade dispute. This is not ideology; it is strategy.

‘Stifled innovation’, ‘censorship’: unpacking the real narrative

The Trump administration justifies its campaign with three arguments that have become second nature: regulation stifles innovation; it favours China (by hindering American companies in their competition with China); and, when it comes to content moderation, regulation is equated with censorship. These narratives are not new, but they are now being promoted with unprecedented intensity and accompanied by concrete means of exerting pressure on economic, diplomatic and informational fronts: threats of tariff barriers against countries that “attack our incredible tech companies”, threats of export controls or restrictions on access to the US market, and travel bans to the United States targeting former European Commissioner Thierry Breton and four representatives of non-governmental organisations fighting disinformation. Added to this are investigations carried out by embassies and coordinated campaigns on social media. As Mathilde Velliet states in her introduction, this study highlights the consolidation of a transatlantic discursive and political ecosystem hostile to digital regulation.

Criticism and statements from members of the administration and Congress regarding European regulations are rife on social media. The €120 million fine imposed by the European Commission on X in December 2025 triggered an immediate surge in activity on the accounts of US officials. The reaction was disproportionate, suggesting that the issue at stake was not the fine itself, but the signal it sent. The United States is fighting back by playing the trade restrictions card – specifically, restrictions on exports of US technology – in an attempt to sway the European position. The Department of Commerce has, moreover, considered new controls on exports of AI chips worldwide, which could require Europeans to obtain the Trump administration’s approval
for their purchases of graphics processing units. As Mathilde Velliet explains, these threats enable Washington to enlist American and European companies in its campaign against regulation, including companies outside the digital sector.

Behind the rhetoric about freedom to innovate, the reality is more prosaic. Generative AI models require massive amounts of data to train. European data – structured, documented and sourced from high-value sectors (health, law, finance) – represents a strategic resource. An effective GDPR, a fully implemented DSA and a binding AI Act: these are all barriers that restrict access to this resource. Dismantling them is therefore not a philosophical struggle for freedom. It is industrial policy. Casting aside his predecessor’s sweeping decree on managing AI-related risks in terms of ethics or security, Trump clearly advocates the idea that “US policy will be to do everything, whatever the cost, to be the world leader in artificial intelligence” (January 2025).

What professionals stand to lose if Europe gives in

Europe is not sitting idly by. But the pressure is real, and it is having an impact. MEPs are echoing the American arguments, some Member States are stalling on the implementation of the AI Act, and there is a temptation to roll back regulation, driven by a section of the business community that sees this as a welcome simplification.

For professionals bound by confidentiality obligations – such as lawyers, chartered accountants, HR managers and healthcare professionals – this debate is far from abstract. The GDPR is not merely a constraint: it is a commitment to their clients. When a law firm or accountancy firm assures its client that their documents are processed in a compliant manner, that assurance is based on a precise legal framework. If that framework cracks under US pressure, what is the promise worth?

This is not a theoretical question. If Europe backs down on algorithmic transparency obligations or data localisation rules, professionals who have built their credibility on compliance will find themselves on shifting ground. The guarantees they offer are only as good as the framework underpinning them.

Regulatory rigour as a competitive advantage

This is where the paradox deserves to be taken seriously. The pressure from the US is intense precisely because the European framework works, at least well enough to be a hindrance. A regulation that changed nothing would not attract so much effort to dismantle it.

There is another way of interpreting this situation: Europe has an advantage that its businesses have not yet fully capitalised on. Sensitive sectors (law, healthcare, human resources, accountancy) will only adopt AI on one condition: the certainty that their clients’ data will not vanish into opaque infrastructures subject to extraterritorial legislation. This certainty is precisely what the European framework is capable of offering, provided it is supported, maintained and effectively enforced. Innovations along these lines must be the first to take centre stage.

In other words, a stable regulatory environment is not an obstacle to the adoption of AI in professional sectors: it should, in fact, be a prerequisite for its adoption. Companies operating in highly constrained environments, subject to strict ethical obligations, do not need fewer rules. They need clear, stable and well-defended rules.

Ifri makes the following recommendation: it is essential to maintain a firm stance, as any retreat would undermine the EU’s credibility and encourage instability. I reiterate my conviction: it is in the EU’s best interests to drive forward efforts towards sovereignty and to encourage innovation that supports this aim and puts sovereignty first.

For organisations that use AI to process confidential data, the same reasoning applies to their own decision-making: regulatory shortcuts do not simplify matters in the long term. They merely shift the risk.

Conclusion

As Mathilde Velliet concludes, Europe must clarify and put into practice its efforts regarding digital sovereignty, starting by mapping out its actual vulnerabilities in the face of accumulated technological dependencies.

But sovereignty is not a matter for a single champion. This is where the debate needs to shift. Concentrating public funding on a single player, however promising they may be, does not drive innovation: it creates dependency under a European flag. Technological competitiveness has never emerged from single-source funding. It arises from an ecosystem where several players, with different approaches, challenge one another, emulate one another and strive to outdo one another. It is competition, not consolidation, that creates the conditions for truly disruptive innovation.

The challenge for Europe is therefore not to replicate Silicon Valley with subsidies, nor to designate its future giants by decree. It is to cultivate fertile ground: a stable regulatory framework, distributed funding, and public procurement markets open to new entrants. Such fertile ground does not yet exist. And until it does, the question of whether Europe can ‘stand up to Trump’ will remain secondary to this one: can it stand up to itself?