Europe’s Digital Sovereignty Hinges on Smarter Regulation for Data Access

Since coming to office, the second Von der Leyen EU Commission has initiated a U-turn in European digital strategy. Within weeks of her second tenure, the Commission withdrew the AI liability directive and more recently, it hastily unveiled a proposal to curb data protection regulation that had previously been Europe’s pride. Shaken by the sudden shift in American foreign policy under United States President Donald Trump, and the use of political power to pressure Europe to abandon its digital rulebook, the EU has awakened to the need for domestic-grown competitors in the technology space.

AI, in particular, has emerged as a domain where Europe wants to avoid missing the boat. As part of this effort, the EU is rapidly scaling back its regulatory aspirations, under the mantra of “simplification.” The current strategy, however, lacks comprehensiveness and ambition, and will ultimately fall short of Europe’s goals by insufficiently addressing the core challenge: incumbent advantage and market concentration in the AI sector. To properly address this, the continent will require bold policy instruments that safeguard its digital sovereignty without compromising its values.

Regulation and Europe’s innovation lag

While political discourse has often been quick to point to onerous European regulation as a reason for the continent’s disadvantage, such a characterization is overly simplistic. Though the EU has more stringent digital regulations compared to the US and others, the alleged lack of competitiveness in this sector predates digital policy efforts.

Furthermore, digital regulation does not differentiate between domestic and foreign companies. In the case of industrial regulation, one can argue that on the global market, more onerous requirements can create a competitive disadvantage. However, for digital technology, the EU can require foreign companies to follow European standards, as is done under the AI Act.

Thus, the argument that Europe lacks a tech industry due to overregulation vis-à-vis international competitors holds little water.

This is not to say that regulation has no impact on market structure. The GDPR, for instance, has actively contributed towards market concentration since compliance is more onerous for start-ups. The key nuance, however, is that these laws have strengthened incumbents, independently of their national origin. It just so happens that established players are US-based.

If no significant European tech industry existed prior to the enactment of the EU’s digital rulebook, and local upstarts are not on unequal footing compared to foreign competitors, watering down tech regulation will not make Europe digitally sovereign.

AI, data & market concentration

Instead, Europe must seek to better understand, and play into, the reality of market competition in the AI sector. Among the factors impacting AI innovation, access to computing power and data are widely recognized as most crucial. While some proposals have been made to address the former, such as making the continent’s supercomputers available to AI start-ups, little has been proposed with regard to addressing the data access challenge.

AI models’ reliance on vast amounts of data favors existing tech giants who have spent decades voraciously accumulating information on every aspect of human existence. Additionally, high-quality data is generally not freely available, with copyrighted books, newspaper articles, and movies ranking amongst the most valuable forms. Large tech companies can afford to negotiate licenses to such content, or plunder it and deal with lawsuits after the fact. Upstarts generally lack this luxury.

Barriers to data access at scale have contributed to the significant concentration of the AI market by making it exceedingly difficult for up-and-coming competitors to gain access to sufficient data to develop a competing product.

For AI start-ups, the challenge is thus threefold. They require access to enough data to compete with entrenched competitors, legal certainty, and, if compensation for content is required, that doing so is easy and affordable.

Addressing this challenge necessitates weighing several interests. The trick is to identify a system that fosters AI innovation and competition through data access and legal certainty, while ensuring that content owners are compensated and thus enabled to continue producing high-quality content.

This matters since, as AI systems become more pervasive, content makers risk being squeezed out, despite having been crucial to developing them in the first place. The replacement of Search Engines by AI, for instance, has already led to a severe decline in revenue for news sites and blogs. This may create a self-defeating spiral where AI systems cannibalize content producers until there is no more data to train on, stifling long-term advancement.

Global approaches

Concerns over Big Tech companies capitalizing on others’ content are not new, nor are attempts at resolving these concerns.

A recent example is news publishers seeing their revenues decline through the advent of large online platforms that benefit from their content. To address this, Australia and others have forced online platforms to negotiate licensing agreements with publishers, with mixed results. While these measures achieved revenue sharing with content producers, they greatly advantage large corporations over small outlets, and rely on platform consolidation, applying exclusively to Facebook and Google. Applying this approach to AI would thus fail to address market concentration.

Then there’s Europe’s approach as inscribed in the AI Act. The regulation exempts commercial research from copyright, though content owners can opt out. The EU Commission’s omnibus proposal would further add an explicit provision for AI training on personal data. Neither of these policies, however, meaningfully facilitates access to high-quality data for upstarts, nor do they improve content owners’ ability to collect revenue from their products.

The absence of a policy solution leaves it up to the courts to determine how to navigate copyright and AI training. The US, for instance, appears to be headed in this direction. The process is likely to be drawn out, and while legal systems are well-versed in determining the rights of the parties, this does not guarantee a desirable or sustainable policy direction in the long term. In the meantime, AI start-ups continue to face legal uncertainty.

Ultimately, none of these approaches succeeds at striking a balance between the rights of copyright holders and promoting AI innovation. Nor do they address the need to proactively favor upstarts against entrenched incumbents.

A European digital commons for AI-training data

This is where European policy innovation comes in.

On the matter of data and copyright, all parties have something to gain from a clear legal framework. AI companies want legal certainty and data access. Content owners want compensation and protection. What if the EU made content more accessible with clear legal provisions and provided a means for copyright holders to garner revenue? Enter the European data commons.

The system would operate as follows: copyright holders submit their works to an EU-managed database, taking responsibility for compliance with certain standards, such as privacy requirements. AI companies could then license this content without negotiating individual agreements and gain access to a centralized database. Contributions are transferred to content owners based on the quantity and quality of data provided, which could be measured through proxies such as the type of works, extent of editing and peer-review involved, or breadth of circulation.

If AI companies don’t contribute, yet still use the data, the EU could levy fines on behalf of copyright owners, leveraging data transparency requirements that are already inscribed in the AI Act. Since tech behemoths have already swallowed virtually everything available on the internet into their models, they would have little choice but to participate.

By applying the requirement to AI developers independently of their provenance, the framework ensures EU competitiveness is not adversely impacted. On the contrary, the approach would enable EU-based AI companies to innovate with legal certainty, avoiding the cost and potential chilling effect of lengthy lawsuits compared to their US competitors. Additionally, by putting the onus on copyright owners to make their content accessible, the framework reduces the burden for AI companies to find (or digitize) training material, which affects small companies most.

Furthermore, the system could leverage contributions from large technology corporations to indirectly subsidize access for European AI start-ups. For instance, fees could be adjusted progressively to the size of a company’s AI model, or be subject to a grace period, enough for start-ups to train their models and become commercially viable.

The model mirrors existing EU proposals for data sharing for health and materials, while integrating compensation for data-owners and market diversification into the framework.

Unlike other approaches, this ensures data access and legal certainty for AI companies and revenue for content owners. Furthermore, it goes beyond minimizing the impact of regulation on start-ups by actively boosting European competitiveness, counteracting market concentration. To be clear, a Data Commons would not in itself solve AI market consolidation, but it would address a crucial barrier to entry.

Regulatory ambition, not surrender

Beyond addressing a core challenge in the AI market, the example of the European Data Commons highlights how government action is not just a zero-sum game between fostering innovation and setting regulatory standards.

By scrapping its digital regulation in the rush to boost the economy and gain digital sovereignty, the EU is surrendering its longtime ambition and ability to shape global technology in its image. The answer to a rapidly changing world is not to surrender, but to enact bold and innovative policy strategies that carefully consider the particular features and challenges of the technology sector to promote the innovation Europe needs to thrive.

Let’s start with data.