Written by Julien Ricciarelli-Bonnal

20 July 2026

Europe Is Forcing Google to Open Android to Rival AI Assistants

On 16 July 2026, the European Commission adopted a decision requiring Google to make eleven Android functions effectively interoperable with competing artificial intelligence services. Behind the technical language of the Digital Markets Act lies a much broader issue: who will control the relationship between users, their smartphones and the applications they rely on?

Until now, Gemini benefited from an advantage that competitors could not easily reproduce. That advantage did not come solely from the quality of the model, but from its integration into the operating system, its ability to be invoked quickly, access the device’s context, interact with certain applications and act directly on phone settings. An external assistant could perform extremely well while still remaining dependent on the permissions, interfaces and limitations defined by Google.

The European decision is designed to reduce that imbalance. It requires third-party services to receive comparable access to several Android capabilities, including voice invocation, contextual data, background execution, interaction with applications, screen automation and certain system controls. Most of the measures are expected to be implemented with Android 18, no later than 1 August 2027.

This opening will not immediately turn Android into a perfectly neutral environment. It does, however, change the nature of the competition. The next battle will no longer focus only on which model produces the best answers, but on which assistant becomes the interface capable of understanding an intention, selecting the right service and carrying out an action without forcing the user to move manually from one application to another.

Google’s Advantage Was Never Based on Gemini Alone

Comparing AI assistants as though they operate under identical conditions ignores the power of distribution. A model may outperform its rivals on certain benchmarks, produce more refined responses or appeal more strongly to advanced users, yet remain disadvantaged if it cannot access the same functions as the assistant integrated by default. On a smartphone, conversational quality is not enough; value also depends on the ability to act.

Gemini can already use Android to interpret the context of a request, open an application, control selected device functions, work with local information and carry out certain actions without requiring a long sequence of manual steps. This advantage becomes decisive as artificial intelligence moves beyond the chatbot and becomes an operational assistant.

The power of an operating system lies precisely in this proximity to everyday use. The company that controls permissions, technical interfaces, background processes and access to device resources can determine which experiences feel seamless and which remain fragmented. A function available through a single request to the native assistant may require several steps when used through a rival, even when that rival understands the instruction perfectly.

Europe is therefore not merely asking Google to allow other applications onto Android. It is requiring the access to be effective enough to support genuine competition. A theoretical opening, undermined by poorer performance, additional delays or more restrictive permissions, would do little to change the market.

The Assistant Is Becoming the Smartphone’s New Gateway

For years, users moved between applications by deciding which tool matched each need. They opened their calendar, email, browser, music service or banking application themselves. AI assistants are trying to reverse that logic: users express an intention, and the system selects the relevant applications, gathers the necessary context and organises the action.

This shift moves value towards whoever receives the initial request. If an assistant can book, summarise, search, write, change a setting, send a message or initiate an action in another application, it is no longer simply one service among many. It becomes the layer through which other services are discovered, selected and used.

The question is therefore no longer limited to which search engine, application or model a person prefers. It is becoming a question of which assistant will gain the right to interpret that person’s intentions. The assistant occupying this position may influence which services are suggested first, which data are used and which journeys become natural. Conversational interfaces could gradually concentrate a level of power previously held by browsers, search engines and app stores.

Opening Android could enable companies such as OpenAI, Anthropic or Mistral AI to offer far more deeply integrated experiences, provided they meet the technical and security requirements. It could also create space for specialised assistants designed around a profession, a language, a regulated sector or a particular standard of confidentiality.

But model quality alone will not decide this competition.

Success will also depend on the ability to build a reliable ecosystem, obtain the necessary permissions, convince developers and create experiences simple enough for users to entrust more context and more actions to a single intermediary.

Greater Access Creates Greater Responsibility

Allowing several assistants to access the context of a smartphone inevitably increases the potential risk surface. Some of the functions involved may concern data stored on the device, sensors, information displayed on screen, notifications, system settings or actions carried out in other applications. Interoperability cannot therefore be reduced to an abstract principle of competition.

The European framework includes qualification mechanisms, transparent conditions and safeguards relating to security, consent and system integrity. Google may still impose certain restrictions when they are justified, provided those restrictions remain objective, precise and non-discriminatory. The real balance will be found at the boundary between legitimate protection and a disguised advantage for the native assistant.

For AI providers, this opening creates a new level of responsibility. Giving an incorrect answer already carries a risk. Taking the wrong action on a device creates another, more immediate one. An error could send a message to the wrong recipient, modify a setting, use inappropriate data or trigger an action that commits the user. As assistants become more capable, traceability, confirmation mechanisms and the ability to reverse an action become increasingly important.

Businesses will also need to approach this development carefully. Professional smartphones contain emails, documents, customer information, internal conversations and access to numerous services. Allowing a third-party assistant to use that context cannot be treated solely as an individual employee choice when the underlying data belong to the organisation.

Before extending these uses, companies will need to clarify their strategic and commercial priorities⁠, identify which assistants are authorised, determine which categories of data may be accessed and specify which actions must still require human approval.

Interoperability expands choice, but it does not remove the responsibility of the provider, the user or the organisation that permits these tools to enter its digital environment.

Google Remains Powerful, but Default Status Will No Longer Be Enough

The European requirement does not erase Google’s advantage. Android remains its operating system, Gemini already benefits from deep integration, and the company’s ecosystem includes search, email, mapping, video, productivity tools and many services used every day. Competitors will still need to persuade users to change assistants, earn their trust and demonstrate enough additional value to justify that choice.

Google also possesses a technical understanding of Android that interoperability rules cannot eliminate. It will continue to develop new functions, on-device models and experiences designed natively around Gemini. The Commission requires future capabilities to remain accessible under comparable conditions, but speed of innovation and quality of execution will still matter.

Even so, the possibility of choosing another assistant under conditions closer to those enjoyed by Gemini may alter incentives across the market. Rival providers will have stronger reasons to invest in Android, developers will be able to design applications for several assistants, and users will be able to compare experiences that extend beyond a simple conversation window.

The market may therefore move from competition between applications to competition between layers of intermediation. Every assistant will try to become the one users call first, the one that knows the most about their context and the one they trust with the greatest number of actions. That position will be extremely valuable because it will partly determine which services remain visible and which are reduced to background execution.

The European decision does not select a winner. It simply prevents the most strategic position from being reserved automatically for the company that already owns the operating system. Its effects will depend on how Google implements the requirements, how effectively competitors use them and how much trust users are willing to place in alternative assistants.

Android is therefore not merely becoming more open. It is becoming the arena for a battle over one of the most strategic interfaces of the coming years: the layer that connects human intention with the digital services capable of carrying it out.

👉 Is your organisation preparing for assistants that can access more data and act across everyday tools? Ricciarelli Partners can help you define rules, responsibilities and safeguards through its AI Business Governance offering⁠.