Apple introduced Siri AI on June 8 [1], a rebuilt voice assistant designed to be more personal and context-aware.

This overhaul represents a strategic pivot for the company as it attempts to close the gap with large-language-model competitors like ChatGPT and Claude. By integrating deeper system awareness, Apple aims to move the assistant from a basic tool to a proactive personal agent.

The new assistant is built on Google’s Gemini model as part of the broader Apple Intelligence framework [1]. This architecture allows Siri AI to understand the content on a user's screen, interact with various apps, and utilize user-specific information to provide more accurate responses [6].

The release follows approximately two years of development delays [4]. The company said this version will place the assistant back at the center of its AI strategy by making interactions more conversational [1].

Alongside the AI updates, Apple announced macOS 27 "Golden Gate" [2]. Reports on the assistant's form factor vary; some sources said it is a rebuilt version integrated directly into the operating system [1], while others said macOS 27 includes a standalone Siri AI app [2].

Siri AI is designed to handle complex requests by leveraging the combined power of on-device processing and cloud-based models. This allows the assistant to maintain a level of personal context that previous iterations lacked, enabling it to perform tasks across different applications without manual navigation by the user [6].

“Siri AI is a rebuilt, more personal voice assistant that leverages Gemini and Apple Intelligence.”

Apple's decision to build Siri AI on Google's Gemini model signals a departure from total in-house development in favor of strategic partnerships to accelerate AI deployment. By prioritizing 'screen awareness' and personal context, Apple is shifting the utility of the voice assistant from a general search tool to a functional OS layer that can execute actions across third-party apps.