EU Digital Markets Act Review Expands to Cover AI Technology Choices on Devices

--- title: "EU Digital Markets Act Review Expands to Cover AI Technology Choices on Devices" slug: eu-dma-review-ai-technology-expansion category: policy story_number: "14" date: 2026-05-06 ---

The European Commission has fired a starting gun on AI regulation through competition law, concluding its first statutory review of the Digital Markets Act with a clear message: artificial intelligence is now squarely within the DMA's enforcement crosshairs.

Published on April 28 as COM(2026) 178, the Commission's review report found the DMA "fit for purpose" and requiring no legislative amendment — but identified AI and cloud computing as priority areas for future enforcement action. The finding transforms what was primarily a regulation targeting app stores, browsers, and messaging services into a framework that could reshape how AI assistants and models are distributed on consumer devices across Europe.

What the Review Actually Says

The Commission's report stops short of formally designating AI services as a new "core platform service" category under the DMA. Instead, it takes a more surgical approach: using existing obligations around interoperability, self-preferencing, and default settings to ensure that gatekeepers cannot lock users into proprietary AI systems bundled with their hardware.

The review highlights that stakeholder consultation revealed overwhelming support for applying DMA principles to AI services. Cloud and AI received more engagement than any other topic during the consultation period, with respondents broadly backing the extension of contestability principles to these emerging services.

Three distinct camps emerged during consultation. The first argued that vertically integrated AI services — such as Apple Intelligence on iPhone or Google's Gemini on Android — already fall under existing DMA obligations, particularly Article 6(7), which requires gatekeepers to allow users to change default settings. The second camp pushed for formal designation of AI assistants as a new core platform service. The third urged caution, arguing that the AI market is too nascent for heavy-handed intervention.

The Commission sided with pragmatism: enforce existing rules aggressively while investigating whether formal designation is warranted by November 2026.

Google Faces Immediate Action

The most concrete enforcement step came even before the review's publication. The Commission opened specification proceedings requiring Google to clarify how it ensures third-party AI developers — including providers competing directly with Gemini — can access Android hardware and software features on equal terms with Google's own services.

The proceedings target Google's control over the Android ecosystem, where Gemini enjoys privileged integration with core system features. Under the DMA's interoperability requirements, rival AI providers like Anthropic's Claude and OpenAI's ChatGPT must be able to access the same device capabilities — microphones, cameras, on-device processing, and system APIs — that Google grants its own AI assistant.

Margrethe Vestager's successor as competition chief, Teresa Ribera, stated in the Commission's announcement that the DMA has "opened up new opportunities for businesses and developers, while giving users more control over their experiences and devices." The Commission's focus on AI interoperability signals that this control must extend to choosing which AI system processes your requests, summarizes your emails, and generates your images.

Apple Moves Preemptively

Apple appears to have read the regulatory wind correctly. Bloomberg reported in March that Apple plans to open its AI stack to rival providers through a system called "Extensions" in iOS 27, allowing users to select between multiple AI models — including Claude and Gemini — for core system features like Siri and Writing Tools.

The timing is revealing. Apple delayed its Apple Intelligence suite in Europe throughout 2025 over DMA interoperability concerns, and only began allowing EU users to set a default voice assistant other than Siri under iOS 26.2 earlier this year. The Extensions framework represents a more comprehensive capitulation: rather than fighting regulators on each specific obligation, Apple is building a platform layer that makes AI providers interchangeable.

The Commission's review effectively validates this approach. By signaling that AI choice on devices falls within the DMA's existing framework, Brussels has created regulatory certainty that rewards proactive compliance over litigation.

Cloud Designations Loom

Beyond device-level AI, the review opens a separate investigation into whether major cloud providers — Amazon Web Services, Microsoft Azure, and Google Cloud — should be designated as gatekeepers under the DMA. This decision, expected by November 2026, could impose interoperability and data portability requirements on cloud-hosted AI services, affecting everything from enterprise AI deployments to API-based model access.

The cloud investigation will examine obstacles to interoperability between services, limited data access for business users, tying and bundling practices, and potentially imbalanced contractual terms. For AI companies that depend on cloud infrastructure — which is effectively all of them — the outcome could determine whether switching between providers remains practically feasible.

Regulatory Competition Heats Up

The EU's move comes as transatlantic tensions over tech regulation intensify. The United States has taken a markedly different approach under the current administration, favoring voluntary commitments over binding regulation. The contrast is stark: while Brussels mandates AI choice through competition law, Washington has largely deferred to market dynamics.

For global technology companies, the EU's approach creates a compliance baseline that often becomes the de facto global standard — the so-called "Brussels Effect." If Apple must offer AI provider choice in Europe, the engineering cost of maintaining separate iOS versions for different markets may push the company toward offering the same choice everywhere.

The DMA review also arrives as the EU AI Act's provisions continue phasing in through 2026, creating a two-pronged regulatory framework: the AI Act governing safety and transparency, the DMA governing market access and contestability. Together, they represent the most comprehensive regulatory architecture for artificial intelligence anywhere in the world.

What Comes Next

The Commission's investigation into AI designations must conclude by May 2027, with cloud designations arriving sooner in November 2026. In the interim, enforcement of existing obligations against designated gatekeepers — Apple, Google, Amazon, Meta, Microsoft, and ByteDance — will intensify, with AI interoperability as a stated priority.

For consumers, the practical outcome is straightforward: European regulators are working to ensure that buying an iPhone or Android device does not mean being locked into a single company's AI ecosystem. Whether that translates into genuine competition or merely a choice screen that most users ignore remains an open question — but after two years of the DMA, Brussels is betting that mandated choice eventually reshapes markets.