European Commission Mandates AI Interoperability on Android Under Digital Markets Act
Key Takeaways
- ▸EC mandates Google enable third-party AI assistants with system-level integration on Android, not just app-level access
- ▸Multiple AI services must be allowed to simultaneously monitor for wake words and run in background with hardware and sensor access
- ▸Third-party AI gains access to Google's on-device models and data from Google apps (Gmail, Calendar, Maps, etc.) with no opt-out
Summary
The European Commission has issued binding specification measures to Google under the Digital Markets Act, requiring the tech giant to enable fair competition for AI services on Android devices. The guidance encompasses two major areas: first, Google must create APIs and system-level integrations that allow third-party AI assistants (such as Claude, ChatGPT, and others) to access the same hardware and software capabilities as its own Gemini AI. Second, Google must share large-scale search data with competitors, including search terms, click data, and usage patterns, at fair and reasonable pricing. The measures aim to level the playing field and prevent Google from leveraging its dominant position in both mobile operating systems and search to unfairly advantage its own AI services.
The Android AI interoperability requirements are particularly sweeping. Third-party AI assistants must be granted access to hardware controls (physical buttons), screen capture from any app, microphone and camera sensors, persistent background operation, and the ability to run audio models on the device's digital signal processor for custom wake-word detection. Most notably, the EC mandates that Google cannot restrict which AI assistants can run concurrently—meaning users can have multiple AI services (Gemini, Claude, ChatGPT, etc.) all monitoring for their own wake phrases simultaneously. Third-party AI must also access Google's on-device local AI models and data from Google's ecosystem apps (Gmail, Calendar, Docs, Maps), with no opt-out for Google and no ability for third-party developers (like Slack) to restrict data sharing to only specific AI providers.
The search data sharing measure requires Google to provide anonymized but granular information about search behavior to competitors and AI services, though search terms themselves present anonymization challenges given their often personally identifying nature. Google is permitted to charge for this data access under Commission-approved fair, reasonable, and non-discriminatory (FRAND) pricing methodology. Together, these measures represent an unprecedented intervention into how a platform operator can leverage its own AI services and data advantage over competitors.
- Google must share search data (terms, clicks, device/language info) with competitors at FRAND-compliant rates
- Guidance prevents app developers from restricting AI data sharing to specific providers—all system AI services must have equal access



