Apple Announces Siri AI Powered by Google Gemini

Apple announced Siri AI, a new conversational assistant powered by next-generation Apple Foundation Models built in collaboration with Google Gemini. It features personal context, screen awareness, system-level app operations, and natural multi-turn dialogue, launching with iOS 27 in 2027.
Key insights
Core Capabilities
Siri AI can cross-application personal context (e.g., find a family-recommended event from iMessage, extract recipe from email, add ingredients to Reminders). It has screen awareness for questions about current content, system-level app operations with third-party apps (WhatsApp, Audible, soon Outlook, Notability), and natural multi-turn dialogue with a dedicated Siri app and iCloud-synced history.
Technical Architecture
The model is next-generation Apple Foundation Models (AFM) custom-built with Google Gemini. Privacy architecture uses on-device processing and Private Cloud Compute (PCC) with no data storage or accessibility, verifiable by external experts. System orchestration uses a system scheduler to call on-device Spotlight index and App Toolbox, prioritizing local data over full model inference.
Future Features and Pricing
Apple Watch Audio Intelligence (late 2026, English beta, Series 12/Ultra 4) uses S11 chip Secure Exclave for audio processing, including sound recognition for hearing-impaired users, Live Rewind, and encrypted summaries. Health app adds Insights and Longevity tabs. Siri AI introduces usage limits and future subscription fees.