Apple announces the opening of Siri for integration with third-party AI assistants, is the iPhone the best distribution layer for AI?

Apple plans to launch the iOS 27 system this year, allowing Siri to integrate multiple third-party AI assistants, breaking the previous exclusive cooperation with ChatGPT, enabling mainstream AI services like Google Gemini and Anthropic Claude to be used directly through Siri. This move is seen as a key step for Apple to catch up with the AI wave and reshape the competitive strength of the iPhone platform.

Siri will support multiple AI services, breaking the exclusive integration with ChatGPT.

According to Bloomberg, Apple is preparing to open Siri for integration with third-party AI assistant services through a new “Extensions” feature in the upcoming iOS 27 system.

This means that users in the future can directly call Google Gemini, Anthropic Claude, Perplexity, Meta AI, xAI Grok, Microsoft Copilot, and other AI services within the Siri interface, breaking the current exclusive collaboration with ChatGPT.

In the future, users can specify a particular AI service for each query and manage all integration switches in the settings page.

WWDC 2025 will officially unveil the details.

Apple is expected to announce specific details at the WWDC Global Developers Conference on June 8, 2025, at which time the settings pages for iOS 27, iPadOS 27, and macOS 27 will introduce a dedicated management interface for Apple Intelligence and Siri, directly guiding users to the App Store’s AI service section to download the required applications.

However, reports also indicate that the related features may still undergo adjustments or delays before their official release.

Creating a one-time integration system to accelerate AI ecosystem expansion.

The strategic significance of this openness lies in Apple being able to break free from the past model of negotiating integration agreements with each AI vendor individually. Once the Extensions system is established, any compliant AI application can connect to Siri through a unified framework, significantly lowering the cooperation threshold and accelerating the proliferation of AI features on Apple devices.

In terms of business model, Apple is also expected to combine the subscription revenue-sharing mechanism of the App Store to gain profit from paying users of third-party AI services, further expanding service revenue.

Google’s stock price fell in response, with the market focusing on the reshaping of the AI landscape.

Following the announcement, Alphabet’s stock fell about 3% that day, closing at $280.74 per share; Apple’s stock remained relatively stable, closing at $252.89. Market movements also reflect high concern about the potential restructuring of the AI ecosystem.

It is worth mentioning that the launch of the Extensions system is different from another ongoing collaboration plan with Google. The latter attempts to rebuild Siri’s underlying technical architecture using the Gemini model through “model distillation,” with both advancing in parallel.

(Apple is distilling a smaller Siri model using Google Gemini, rewriting the intelligent assistant in iOS 27)

Success or failure? It depends on execution quality.

Throughout history, Apple’s true advantage has never been the capability of AI models, but rather device integration and user habits. If the Extensions system is executed smoothly, allowing users to call Claude or Gemini within Siri as naturally as switching songs, Apple can indeed maintain control over the entry point of AI usage without winning the AI competition. However, it is also true that Siri has seen technological stagnation over the past decade.

Opening access to external AI models could be a genius move for Apple to create real-world AI applications, or it could be an admission of failure in advancing its own models. The success of this decision will entirely depend on the quality of execution and integration.

This article “Apple announces open integration of Siri with third-party AI assistants, is the iPhone becoming the best distribution layer for AI?” first appeared in Chain News ABMedia.