On July 16, the European Commission dropped a regulatory bombshell: Google must let rival AI assistants like ChatGPT, Claude, or Perplexity work as deeply inside Android as its own Gemini does. The binding Digital Markets Act decision gives Google until July 2027 to implement changes, with search data sharing starting January 2027. For Windows users, the ruling reads as a preview of what regulators might demand from Microsoft’s own AI ambitions with Copilot.

What the EU Actually Ordered Google to Do

The Commission’s Android decision covers 11 operating-system capabilities it calls the building blocks for capable AI services. These include:

  • Invocation methods: long-press home, navigation gestures, and always-listening hotwords, even when the screen is off or the device is in battery saver.
  • Contextual data access: user-approved reading of screen contents, notifications, app data, and sensor signals.
  • Actions across apps and the OS: using Android’s App Functions and Computer Control automation to draft emails, create calendar entries, order groceries, or book travel.
  • On-device AI models: comparable access to Gemini Nano and other system-level models for offline tasks like transcription and summarization.
  • Background execution: equal ability to run background tasks without being prematurely killed.
  • Microphone, camera, speaker, and location access under the same consent and awareness conditions applied to Google’s own services.

A second decision compels Google to share anonymized search data with eligible search engines and AI chatbots that offer search features. The Commission rejected Google’s earlier proposal because it stripped out 90–100% of unique queries, rendering the data useless. The new framework requires multilayer anonymization with privacy expertise, aligned with draft guidance from the European Data Protection Board. Google may charge fair, reasonable fees to cover incremental costs.

The deadlines are clear: January 2027 for search data sharing; July 2027 (no later than August 1) for the Android changes, which must ship with the next major Android release. These are not fines but specification decisions—failure to comply opens the door to enforcement proceedings, including potential fines of up to 10% of global turnover.

Why This Decision Matters Far Beyond Android

For EU Android users, the payoff could be real choice. Today, even if you set a third-party assistant as the default, it lacks the deep system integration that makes Gemini feel seamless—no always-listening hotword when the screen is off, no ability to read contextual data or automate multi-step workflows. By mid-2027, you should be able to wake your preferred assistant by voice from a locked phone and have it complete a task like sending a WhatsApp message or adding a calendar event without manually opening each app.

But the ripple effects extend to every platform with an integrated AI assistant—and Windows is Exhibit A. Microsoft has woven Copilot into the taskbar, Edge, Office, and soon deeper into Windows through “Windows App Actions” and local AI models on Copilot+ PCs. If the DMA logic holds, the EU could eventually require Microsoft to provide similar access to rivals like ChatGPT, Google Gemini, or others on Windows. Apple is already under DMA pressure to open iOS and iPadOS interoperability. The Android decision sets a concrete template: define the essential AI-enabling OS capabilities and demand fair access.

Google argues the order threatens security and privacy, as Kent Walker emphasized. The Commission counters that its specifications include consent requirements, data integrity checks, and the ability for Google to assess risks before sharing data. For Windows users, the debate is familiar: Microsoft has similarly warned that forcing open its AI integrations could erode safety. The EU’s approach suggests those arguments won’t automatically prevail—platforms will have to prove that parity isn’t possible without undue risk.

How We Got Here: The DMA’s Long Reach

The Digital Markets Act, in effect since 2022, designated Google (and Alphabet) as a gatekeeper for multiple core platform services, including Android and Google Search. Previous DMA actions pushed Apple and Google to open app stores and messaging. The July 2026 specifications build on a 2025 investigation into Google’s AI practices. The Commission concluded that Google’s control over Android’s AI stack—from invocation to on-device models—hindered the emergence of competing AI assistants, a violation of the DMA’s interoperability obligations.

Parallel cases are unfolding: the Commission has been probing Apple’s iOS restrictions, and it has reportedly been examining Microsoft’s practices around Windows and Bing. In January 2026, the EU sent Microsoft a preliminary finding about Edge and Bing integration, but no AI-specific action has yet been taken. The Android ruling may accelerate similar scrutiny of how Windows treats Copilot versus other assistants.

For Developers: A More Level Playing Field

Developers building AI assistants should start planning for Android as a genuinely open platform. By mid-2027, you could ship an app that users invoke with “Hey Assistant” from a locked screen, reads contextual data with consent, and executes multi-step workflows using Android’s automation APIs. The Commission’s specifications require that any enhancements Google makes to these capabilities for Gemini must be made available to third parties simultaneously. That means you can innovate on a more equal footing—no more building a brilliant conversational AI that remains a second-class citizen because it can’t interact with the screen or run in the background like Gemini.

The search data decision is equally important for AI companies. Access to high-quality, anonymized query logs can dramatically improve relevance, ranking, and retrieval for AI answer engines. By January 2027, eligible services can request this data under fair terms, which could level the field for newcomers trying to build competitive search experiences.

For IT Admins: A Wake-Up Call for Windows

For IT administrators managing Windows fleets, the Android decision is a wake-up call. If EU regulators apply the same framework to Windows, corporate devices could see a similar mandate: Microsoft might have to let users install and deeply integrate alternative AI assistants. That raises complex permission management, data leakage risks, and compliance challenges.

IT teams should begin evaluating how they would handle a multi-assistant environment—policy controls, app allowlists, EMM extensions, and security auditing for AI actions. On Android, the Commission stressed that users must explicitly authorize each assistant and that apps can limit what data is shared. Windows IT can expect comparable guardrails, but managing them at scale won’t be trivial. Start piloting Microsoft Intune’s app protection policies for AI assistants and demand that vendors provide detailed permission models and audit logs.

What Windows Users Should Watch Now

While the immediate order targets Android, Windows users should note three signposts:

  1. Copilot’s evolving integration: Microsoft is embedding Copilot deeper into Windows 11 and the upcoming Windows 12. The more it uses system-level access (screen analysis, app control, local AI models), the more it resembles the capabilities the EU just regulated on Android.
  2. EU statements on Microsoft: The European Commission has not yet issued a similar specification for Windows, but it is actively monitoring the platform’s AI integration. Any formal investigation would follow a similar pattern: define gatekeeper status for Windows, identify anti-competitive AI advantages, and demand interoperability.
  3. User choice as default: Even without regulation, the trend is toward user choice. Windows already lets you change the default browser and search engine; an AI assistant could be next. Microsoft’s own messaging has evolved to emphasize partner integrations (like Meta AI on Bing), suggesting a possible voluntary opening.

Action Items: Preparing for an Open AI Assistant World

  • Users: If you’re in the EU, mark July 2027 on your calendar. You may soon be able to replace Gemini with a preferred assistant on Android with full functionality. On Windows, keep an eye on upcoming builds; Copilot’s tight integration could become more modular if Microsoft anticipates regulation.
  • Developers: Study the Commission’s list of 11 OS capabilities. Design your AI assistant to leverage them. Start discussions with OEMs about pre-installation and with Google about access terms.
  • IT admins: Begin piloting Android and Windows AI management tools. Explore Microsoft Intune’s app protection policies for AI assistants and demand that vendors provide detailed permission models and audit logs.
  • Everyone: Watch for Google’s implementation proposal, expected within months. It will clarify how much real access competitors get. Also follow the EU’s ongoing DMA cases against Apple and any formal steps toward Windows.

The Outlook: Android as a Test Bed for Windows

The Android decision establishes that a platform’s integration of its own AI assistant—using exclusive access to OS hooks, background execution, and on-device models—can be an antitrust problem. Microsoft’s Copilot strategy, especially on Copilot+ PCs with dedicated NPUs, looks remarkably similar. While Microsoft isn’t yet a DMA gatekeeper for AI-specific practices on Windows, the Commission has signaled it won’t ignore how AI is built into the OS.

If a future EU order forces Windows to open up, it could mean that users can set ChatGPT or Gemini as the default assistant that responds to the Copilot key, accesses local models, and controls apps through Windows App Actions. Microsoft would likely resist, citing security, but the Android precedent suggests that regulators will demand concrete evidence of risk before allowing closed systems.

For now, Android is the test bed. Google has until 2027 to comply while preserving security. How well it walks that line will influence what Brussels demands of Microsoft, Apple, and others. For Windows users, the takeaway is simple: the phone in your pocket is about to gain a more diverse set of AI brains, and your PC might not be far behind.