Apple Can Create Smaller On-Device AI Models From Google’s Gemini

Apple has full access to Google’s Gemini model, allowing them to customize it for Siri and other AI features on their devices. According to The Information, Apple can use this access to create smaller models that are optimized for specific tasks and designed to run efficiently on Apple devices without internet connectivity.

Distillation Process

The process involves feeding high-quality results from the main Gemini model into a smaller model. This enables the smaller model to learn the internal computations used by Gemini, resulting in efficient models that perform like Gemini but require less computing power.

Customization and Challenges

Apple can edit Gemini as needed to ensure it responds to queries in a way that aligns with Apple’s vision. However, they face challenges because Gemini is tuned for chatbot and coding applications, which doesn’t always meet Apple’s needs.

Siri Enhancements

Apple plans to introduce a smarter, chatbot version of Siri on iOS 27, leveraging Google’s Gemini models. This enhanced Siri will be capable of answering questions, summarizing information, scanning and understanding uploaded documents, providing emotional support, and completing tasks like booking travel.


元記事: https://www.macrumors.com/2026/03/25/apple-google-gemini-distill-models/