Apple announces Siri AI powered by Google Gemini models at WWDC 2026

Apple announced Siri AI at WWDC 2026, revealing a "deep collaboration with Google" that leverages technologies behind the Gemini family of models. According to Apple, this partnership resulted in the "next generation of Apple Foundation models" for Apple Intelligence.

Architecture and capabilities

Siri AI represents what Apple describes as a transformation comparable to Google's transition from Assistant to the Gemini app. The new system includes personal context understanding, app actions, on-screen awareness, image understanding, and access to broad world knowledge with conversational back-and-forth capabilities.

The interface uses the Dynamic Island, with answers expanding from the top of the screen. Users can swipe down to view full responses, access the system via long-press on the side button, or use the "Hey Siri" hotword. Apple added a dedicated Siri app displaying conversation history in a card-based feed.

Voice and dictation improvements

Apple announced a more expressive Siri voice with multiple options and adjustable sliders for Pace and Expressivity. Dictation accuracy received what Apple calls a "major boost" with improved spelling, punctuation, and capitalization.

Camera integration

Siri AI adds a new mode to the Camera app offering guided experiences. According to Apple, users can point their iPhone at food for nutritional insights, split restaurant bills by selecting ordered items, or add multiple events to Calendar from a photographed schedule.

Safari enhancements

Safari uses Apple Intelligence to organize tabs into topics and includes a "Notify Me" feature for page monitoring. The browser can navigate through websites to automatically fix passwords. A "Describe an extension" feature creates custom extensions from natural language descriptions.

Cross-app integration

Messages offers one-tap suggestions for creating reminders or notes based on conversation context. Calendar accepts conversational event creation. The Phone app surfaces relevant information like airline confirmation codes during calls. The Home app combines multiple alerts into single notifications and can summarize camera clips.

Image Playground now generates photorealistic images, though Apple did not disclose the underlying model. Photos uses "more powerful image models" for Clean Up tool improvements and a new Extend feature. Spatial Reframing allows users to change shot perspectives as if repositioning the camera in the original scene.

Subscription model

Like Google One, iCloud+ will provide increased access to features. Apple stated that image generation has daily usage limits "because they rely on powerful server models," with subscription tiers offering higher limits. Specific pricing was not disclosed.

What this means

Apple's adoption of Google's Gemini models marks a significant shift in its AI strategy, moving away from complete in-house development. This partnership gives Apple immediate access to competitive AI capabilities while Google gains distribution across Apple's ecosystem. The integration appears deep enough that Apple is building its foundation models on top of Gemini technology rather than simply offering Gemini as an optional assistant. The subscription model for compute-intensive features mirrors industry trends toward tiered AI access based on server resource consumption.