Gemini technology won’t be directly built into Siri AI

Revamped digital personal assistant AI models were developed and optimised by Apple

Life

Although Apple is working with Google on the development of its new Siri AI, that does not mean the voice assistant simply runs on Google’s own Gemini model. Apple emphasises that Siri AI uses its own AI models and that no Gemini software or Google search functionality is directly embedded in the assistant. The cooperation is limited to underlying model development and training methods.

During the development of Apple Intelligence, Apple developed part of its Foundation Models together with Google. In doing so, as has now become clear, it made use of technology from the Gemini models and of training methods in which the outputs of advanced Gemini models were used to further improve Apple’s own models. However, the final AI models were developed and optimised by Apple itself.

According to Apple, Siri AI consists of a set of foundation models that can run both locally on compatible devices and in the cloud. For simple tasks, processing takes place on the device itself. More complex commands are handled via Apple’s own cloud infrastructure. Only for the most computationally intensive tasks can Google infrastructure be used, with Apple stating that privacy protection is preserved.

Apple also distinguishes between the technology behind an AI model and the user experience. According to the company, Google supplies part of the technical foundation, while Apple remains responsible for integration with iOS, operation, privacy measures, linking with apps, and the way in which Siri generates answers. As a result, Siri remains, according to Apple, a product in its own right and not a version of Gemini with a different interface.

Emerce

Read More: AI Apple Artificial Intelligence digital personal assistants Gemini Google Siri