Siri AI to Use Google Gemini, Focus on Deep App Integration

Apple's move to integrate Google's Gemini LLM into Siri AI underscores a pragmatic partnership despite their competitive relationship. The architecture spans four layers—devices, models, reasoning, and applications—with on-device AFM handling low-latency private data tasks, while cloud AFM focuses on long-form understanding and complex reasoning. This division allows Apple to maintain privacy through private cloud computing while leveraging Gemini's capabilities.

Siri AI's broader device coverage across iPhone, MacBook, and Vision Pro gives it a footprint advantage over rivals like Google, Samsung, and Xiaomi. However, the report notes that its orchestrator architecture faces challenges in maturity, ecosystem, and service gaps, and the integration of third-party apps is still immature with no killer application yet. Apple provides development frameworks and tools to expand the Siri AI ecosystem, aiming to strengthen its cross-app integration advantage in the future.