Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Brussels Draws a Line in the Android Sand
The European Commission has issued a landmark interoperability order requiring Google to open up core Android capabilities — the microphone, camera, on-screen content, an always-on wake word, and the ability to control other apps through simulated taps and text input — to third-party AI assistants on equal terms with its own Gemini. The ruling, handed down on Thursday, sets a hard deadline: full compliance must ship with Android 18 and be live no later than 1 August 2027.
The decision is significant not just for its technical scope but for what it reveals about how European regulators now think about AI competition. Rather than targeting search rankings or app store fees — the familiar battlegrounds of past antitrust fights — Brussels is going straight for the sensory infrastructure of the smartphone: the eyes, ears, and hands that allow an AI assistant to be genuinely useful rather than merely present.
What “Equal Access” Actually Means
Currently, Gemini benefits from deep system-level integration that competitors cannot replicate. When a user asks Gemini to describe what is on their screen, book a restaurant by calling up a third-party app, or respond to a notification without unlocking the phone, it can do so because Google has wired those hooks directly into Android’s architecture. A rival assistant — whether from Anthropic, OpenAI, Samsung, or a European startup — must make do with whatever public APIs Google chooses to expose, which is a considerably narrower set of tools.
The Commission’s order effectively treats those privileged hooks as essential infrastructure. Under the new rules, any assistant that meets Google’s technical and security standards must be able to invoke the wake word while the screen is off, read live camera and microphone streams, interpret the current on-screen context, and perform in-app actions in the background. In other words, the full agentic toolkit — not a watered-down approximation of it.
The Android 18 Timeline
Tying compliance to Android 18’s release cycle is a deliberate choice by regulators. It prevents Google from stretching the deadline through bureaucratic delay while still giving the company a realistic engineering window. The August 2027 backstop ensures that even if Android 18 slips, the obligation does not. Google has not publicly disputed the timeline, though it has historically challenged the substance of European competition findings in court, and a legal appeal remains a live possibility.
Practically, engineers at Google will need to design an API surface broad enough to satisfy the order while robust enough to prevent misuse — a non-trivial task when the capabilities in question include continuous microphone access and arbitrary in-app automation. Security and privacy guardrails will need to be standardised, since any weakness in a third-party assistant’s integration becomes a weakness in the Android ecosystem as a whole.
Winners, Losers, and the Bigger Picture
The immediate beneficiaries are the assistant makers who have long complained that competing on Android feels like playing an away game on a pitch Google owns. Samsung’s Bixby, which already runs on Android hardware but struggles to match Gemini’s integration depth, could see a meaningful upgrade path. European AI ventures building assistant products have even more to gain, given that they currently face both a capability gap and a home-market regulatory environment that is now, in principle, working in their favour.
For users, more genuine competition among AI assistants on the world’s most widely used mobile operating system could translate into faster innovation, better privacy options, and alternatives that do not feed data back into Google’s advertising machine.
Analysis
This ruling signals a maturation in how regulators approach AI market power. Earlier interventions focused on distribution — which app is pre-installed, which search engine is default. The Commission is now targeting capability parity, recognising that in an agentic AI world, access to sensors and system APIs is the new gatekeeping mechanism. Whether Google can design open interfaces that satisfy regulators without undermining its own competitive position will be one of the defining platform battles of the next two years — and the answer will shape how AI assistants evolve everywhere, not just in Europe.
