Europe is moving beyond broad AI principles to enforce concrete rules that reshape how tech companies compete. The European Commission has published new codes of practice on AI-generated content labeling and is actively enforcing data-sharing requirements against dominant search platforms, signaling a shift from theoretical frameworks to practical enforcement that will ripple across the continent's digital economy.

What Is the EU Actually Doing to Regulate AI Right Now?

The European Commission published a final Code of Practice on marking and labeling AI-generated content on June 10, 2026, establishing standards for how companies must disclose when content is created by artificial intelligence. This follows years of foundational work by the Commission's high-level expert group on artificial intelligence, which developed ethics guidelines and practical assessment tools that have guided European AI policy since 2019.

Beyond labeling, the Commission is enforcing the Digital Markets Act (DMA), a competition law that requires dominant platforms to share data with rivals. Google is currently facing specification proceedings for allegedly failing to comply with Article 6(11) of the DMA, which mandates that the company share ranking, query, click, and view data with competitors. This enforcement action represents one of the most aggressive data-sharing remedies in modern tech regulation.

How Does the EU's Data-Sharing Rule Actually Work?

The Commission's preliminary measures require Google to disclose comprehensive anonymized information about search results and user interactions with them. For example, this includes the content of key snippets displayed in response to a search query and how long a user's cursor hovered over a link before clicking. The framework also appears to cover AI-generated features like AI Overviews, which increasingly dominate Google Search's results page, though the Commission should clarify whether AI-related features are explicitly included in the definition of "visual content".

The data-sharing requirement operates on a parity principle: competitors must receive access to search data on the same terms and timeline that Google uses internally. The Commission's preliminary measures specify that Google must share its search data daily, though the interaction between this daily requirement and the broader parity principle needs clarification in the final rules.

Steps to Understanding Europe's AI Governance Framework

• Ethics-First Foundation: The Commission's approach begins with seven key requirements for trustworthy AI, including human agency, technical robustness, privacy protection, transparency, and accountability. These principles, developed by the high-level expert group, guide all subsequent legislative steps.
• Practical Implementation Tools: The Assessment List for Trustworthy AI (ALTAI) translates abstract ethics guidelines into a dynamic self-assessment checklist that developers and deployers can use to implement trustworthiness requirements in real products.
• Enforcement Through Competition Law: Rather than relying solely on AI-specific regulations, the Commission uses existing competition law to force dominant platforms to share data, unlocking innovation by rivals and protecting user interests simultaneously.
• Sectoral Customization: The Commission recognizes that AI governance cannot be one-size-fits-all. It has developed specific considerations for implementation in public sector, healthcare, and manufacturing applications.

This multi-layered approach reflects a maturation of European AI policy. Rather than issuing broad prohibitions, the Commission is now specifying exactly how companies must operate, what data they must share, and how they must label AI-generated content.

Why Does Data Sharing Matter for AI Competition?

The Commission's data-sharing remedy draws inspiration from historical precedent. In 1956, AT&T agreed to license its complete patent portfolio, including transistor technology, royalty-free to competitors. Empirical research found that this mandatory sharing increased innovation by 12 percent relative to non-licensed patents in the same technology class and drove substantial overall innovation growth. The Commission believes search data could play a similar catalytic role in the AI era.

Machine learning models that power search products typically improve when search engines have access to fresh queries, a finding confirmed in US judicial proceedings and independent empirical research. By requiring Google to share daily search data with rivals, the Commission aims to level the playing field and enable competitors to build better search products powered by AI.

However, the scope of data sharing raises important questions. The Commission's approach appears broader than the remedy in the US v. Google antitrust case, which excludes user data that Google uses to train search-related generative AI models. The Commission's preliminary measures appear to include AI features like AI Overviews in the data-sharing requirement, though this needs explicit clarification.

What Privacy Safeguards Are Built Into the EU's Data-Sharing Rules?

Given the scale and sensitivity of search data, privacy protection is central to the Commission's framework. The preliminary measures include comprehensive technical protections intended to reduce the risk of user re-identification to a "residual level". These protections are thoughtful and innovative, but the Commission did not derive them from clearly specified design goals.

The Commission should establish explicit privacy and utility goals to govern the design of anonymization measures and conduct ongoing evaluation by independent experts, according to analysis from the Knight-Georgetown Institute. As search behavior changes and more competitors gain access to shared data, performance against privacy and utility goals may degrade. User behavior is not static, especially as AI-enabled features spark changes in the types of queries received by search products, implying a need for regular reevaluation of how technical measures are designed.

The Commission should also explain the motivation and rationale for each parameter specified in the technical measures, including why it chose each specific protection and what trade-offs it made between user privacy and data utility. Without this transparency, stakeholders cannot assess whether the framework adequately protects users while enabling meaningful competition.

What Comes Next for European AI Regulation?

The Commission's work on data-sharing rules and AI-generated content labeling represents a critical test of whether Europe can regulate AI effectively without stifling innovation. The high-level expert group's original recommendations emphasized that trustworthy AI should empower, benefit, and protect European citizens while driving sustainability, growth, competitiveness, and inclusion.

The European AI Alliance, an online forum with over 4,000 members representing academia, business, civil society, and policymakers, continues to provide feedback on the Commission's approach. This community-driven process suggests that European AI governance will remain iterative, adapting as technology evolves and as the Commission learns from enforcement actions against dominant platforms.

The stakes are high. If the Commission's data-sharing rules succeed in catalyzing competition in AI-powered search while protecting user privacy, they could become a model for other sectors and other jurisdictions. If they fail to balance these competing interests, they risk either stifling innovation or leaving users vulnerable to privacy breaches. The Commission's final measures, expected in the coming months, will determine whether Europe's AI rulebook translates principles into practice.

" }