European regulators have ordered Google to open the deepest layers of its Android operating system to rival artificial-intelligence assistants, a move that could give tools such as ChatGPT, Claude, and Perplexity the same reach into phones that Google reserves for its own Gemini assistant. The binding decisions, issued under the European Union’s Digital Markets Act, set staggered deadlines beginning in 2027 and mark one of the most far-reaching regulatory interventions yet into how AI software interacts with mobile devices.
The orders apply only within the European Union, but their implications for how assistants compete on smartphones extend well beyond the bloc. At stake is control of the assistant layer on the world’s most widely used mobile platform, where deep system access increasingly separates a genuinely useful helper from a peripheral app.
What the Digital Markets Act requires
The European Commission issued its specification decisions in July 2026, formally directing Google to grant competing AI assistants access to a set of Android capabilities that had been effectively walled off for rivals. According to an industry roundup from MarketingProfs, the order covers 11 operating-system features, including the camera, the microphone, the contents of the screen, always-on wake-word detection, and the ability to simulate taps and typing so an assistant can control other apps on a user’s behalf. In practical terms, those are the hooks that let an assistant see what is on the display, listen for a spoken command at any moment, and take actions inside unrelated applications, the kind of system-level integration that distinguishes a built-in assistant from a mere app.
Why this levels the playing field with Gemini
The core complaint the decisions address is one of parity. Google’s own Gemini assistant, as a native part of Android, has been able to draw on system-level capabilities that third-party assistants could not match, giving it a structural advantage on the world’s most widely used mobile platform. By requiring Google to extend comparable access to rivals, regulators aim to ensure that a user who prefers a competing assistant is not stuck with a hobbled version that cannot see the screen, respond to a wake word, or operate other apps. As The Next Web reported, the intent is to prevent the company that controls the operating system from also locking in an advantage for its own assistant at the layer where integration matters most.
The timeline Google must meet
The decisions do not take effect all at once. Google must implement the unrestricted feature set in Android 18 by August 1, 2027, and then follow with concurrent multi-assistant hotword detection, the ability for more than one assistant to listen for its own wake word at the same time, in Android 19 by August 1, 2028. A separate strand of the order addresses search data: anonymized query, click, and ranking information is set to begin flowing to rivals in January 2027, under a method that enforces a minimum cohort size of 1,000 users per shared record to protect individual privacy. That phased schedule, laid out in coverage by Tech Times, gives the company time to re-engineer parts of Android while committing it to hard compliance dates.
The privacy and security questions
Handing outside assistants access to the camera, microphone, screen contents, and the ability to control other apps raises obvious security and privacy stakes. The same capabilities that make an assistant genuinely useful, seeing what is on the screen and acting within other applications, are also powerful surveillance and automation tools if misused. The order’s design attempts to balance competition against those risks, for example by anonymizing shared search data and capping how granular it can be. But the broader tension remains: opening system-level hooks to more companies widens the number of parties that can reach sensitive device functions, and regulators, Google, and the assistant makers will have to demonstrate that the expanded access is governed by strong permissions and safeguards rather than simply multiplied across more vendors.
An EU rule with global ripples
The decisions apply only to Android devices and users in the European Union, so phones in the United States and other markets are not directly affected by the mandate. Even so, rules of this kind have a history of influencing product design far beyond the borders where they apply, because building one version of an operating system for Europe and another for the rest of the world is costly and complex. If Google engineers deeper assistant access into Android to satisfy the EU, elements of that architecture could surface elsewhere, and rival assistant makers that gain a foothold on European devices may push for similar treatment in other jurisdictions. The Digital Markets Act has already reshaped app-store and messaging practices on the continent, and this expansion into AI assistants signals that regulators view the assistant layer as the next competitive battleground.
What comes next
For now, the orders establish a clear direction and a firm set of deadlines rather than an immediate change on any user’s phone. The first concrete milestones arrive in 2027, with search-data sharing in January and the core Android access requirement by August, followed by the multi-assistant wake-word capability in 2028. Between now and then, the practical questions are whether Google implements the changes to regulators’ satisfaction, how the assistant makers take advantage of the new access, and how well the privacy safeguards hold up once multiple assistants can reach into the heart of the operating system. The decisions have set the terms; the coming two years will show how deeply they reshape the way AI assistants live on Android.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
More from Morning Overview
