EU Android DMA rules force Google to open AI assistant stack

The EU is quietly trying to pry open what may be the most important choke point in the post chat era: the operating system layer where AI agents live. For now, Gemini on Android and Siri on iOS enjoy deep privileges that third party assistants cannot touch, from always on voice wake words to cross app automation and full screen access. For an AI that is supposed to act as a true agent, those capabilities are everything. Forcing Google to expose 11 core Android functions to qualified third party AI services, and to share anonymised search data, is an attempt to stop Gemini from becoming the de facto gatekeeper for mobile AI.

If this survives legal challenges, it changes the terrain for OpenAI, Anthropic, Mistral and any European assistant maker. Instead of being trapped inside a browser or a single app, they could offer near native experiences on hundreds of millions of Android devices, with much richer context and control. That could accelerate the shift from chatbots to agentic helpers that manage workflows across apps, calendars and payments. At the same time, it pushes AI competition away from pure model quality into the messy world of interoperability, security certification and platform governance, areas where regulators rather than labs wield more leverage.