EU tells Google to open up Android AI; Google calls it ‘unwarranted intervention’

Back in January, the European Commission kicked off what it calls a “specification proceeding” — basically a formal investigation — into how Google bakes AI into Android. Now the results are in, and surprise: the EU thinks Android needs to be more open. Google, equally predictably, calls it “unwarranted intervention.” Both sides are playing their parts.

This is all happening under Europe’s Digital Markets Act, or DMA, which labels seven big tech companies as “gatekeepers” and subjects them to extra rules meant to keep competition fair. Google has been fighting these regulations for years, but the DMA isn’t going anywhere. The commission has already forced changes to search results, app stores, and messaging interoperability. Now it’s AI’s turn.

The core issue is simple: Gemini ships with every Google-powered Android phone. It’s there from the first boot, deeply integrated at the system level, and gets privileges no third-party AI service can touch. Want to invoke an assistant with a long press of the power button? That’s Gemini. Want to access AI features from the quick settings panel? That’s Gemini too. The commission argues that too many Android experiences are locked to Google’s own AI, and as a gatekeeper, Google has to open those up.

I’ve been watching this play out for a while, and honestly, it’s hard to argue with the EU’s logic here. The DMA was designed precisely for this kind of scenario — a platform holder using its control over the OS to favor its own services. Google has done this before with search, with Chrome, with Google Pay. Now it’s doing it with AI. The fact that Gemini gets system-level hooks that no competitor can replicate is exactly the kind of advantage the law targets.

That said, Google’s pushback isn’t entirely baseless. Deep AI integration isn’t just a marketing gimmick; it requires tight coupling with the OS for performance, privacy, and security. Opening that up to any third party could mean more attack surface, worse battery life, and a fragmented user experience. Google’s argument that this is “unwarranted intervention” into product design has some technical merit, even if it’s self-serving.

But the commission isn’t asking for everything. They’re not demanding Google hand over its model weights or open-source Gemini. They want parity in access points — the ability for a user to set a default AI assistant, for third-party AI services to register for system-level triggers, and for Google to stop preloading Gemini in ways that can’t be removed or replaced. That feels reasonable, especially since similar rules already apply to browsers and search engines on Android under the DMA.

If the commission moves forward — and I expect they will — we could see changes as early as this summer. Google would likely have to offer an AI assistant choice screen on new Android phones in Europe, similar to the browser choice screen we’ve seen for years. Third-party AI services could register for hardware buttons, voice wake-up, and maybe even on-device API access.

The bigger question is whether this sets a precedent. The US, UK, Japan, and others are all watching how the DMA plays out. If the EU forces Google to open up Android AI, it could become a global standard — or at least inspire similar actions elsewhere. Google knows that, which is why they’re fighting harder than usual.

For now, we wait. The commission will likely issue a formal decision in the coming months. Google will appeal. Lawyers will get rich. And Android users in Europe might eventually get to choose which AI runs their phone. That’s a win for choice, even if it makes Google’s engineers grumble.