With iOS 27, Apple may open Siri and AI features to third-party models

Apple is reportedly planning to open Apple Intelligence to third-party AI services like Google Gemini and Anthropic Claude with a new 'Extensions' system in iOS 27

With iOS 27, Apple is reportedly planning to open its Apple Intelligence ecosystem to rival services from companies such as Google and Anthropic. According to a Bloomberg report, Apple will allow users to select different AI models for tasks such as writing, image generation, and editing. This will expand on Apple’s existing partnership with OpenAI, which currently integrates ChatGPT into Apple Intelligence.

iOS 27: How the new system could work

Inside iOS 27, Apple is reportedly referring to the feature as “Extensions”. It will allow users to choose which AI service powers different features directly from the Settings app. These AI models could be used across key functions such as Siri, Writing Tools, and Image Playground, enabling users to route tasks to different services depending on their needs.

 

Apple is also expected to introduce a dedicated section in the App Store to highlight compatible AI apps.

 

The update is also expected to bring deeper changes to Siri. Users may be able to replace ChatGPT with other AI chatbots within Siri, effectively turning the voice assistant into a gateway for multiple AI services.

Another reported change includes the ability to use different voices depending on which AI model is responding. For example, responses generated by Apple’s own system could sound different from those handled by third-party tools such as Anthropic’s Claude.

 

However, the report noted that Apple plans to clarify it will not be responsible for content generated by third-party AI models, even as it continues offering its own in-house AI capabilities alongside external options.

Other expected iOS 27 additions

Other reported additions include AI-powered photo editing tools, a Siri mode within the camera app, and new ways to create custom passes in the Wallet app.

 

Apple is also planning to add a new Siri option alongside existing camera modes such as Photo and Video. This new mode is expected to replace the current standalone Visual Intelligence experience and integrate it directly into the main camera interface.

 

In terms of design, iOS 27 is not expected to bring a major visual overhaul, though Apple may continue refining the Liquid Glass interface introduced with iOS 26.

Google-Apple collaboration

In January, Apple and Google announced a multi-year partnership under which the next generation of Apple Foundation Models will be based on Google’s Gemini models. These models will power future Apple Intelligence features, including a more personalised Siri expected later this year.

 

Recently, Google Cloud chief Thomas Kurian, speaking at Google Cloud Next 2026 on April 22, said Apple remains on track and that the revamped Siri is expected to arrive in 2026 itself.

 

For the uninitiated, Apple outlined its upgraded Siri vision at WWDC 2024, focusing on a more integrated and intelligent assistant experience.

 

The new version is expected to bring contextual awareness and cross-app functionality, enabling Siri to understand on-screen activity and execute multi-app tasks through natural commands.

 

It may also introduce more conversational voice interaction and multimodal capabilities, allowing Siri to process text, voice, and visual inputs together.

EU pressure on AI openness

On April 27, the European Commission issued preliminary findings asking Google to open key Android features to third-party AI services under the Digital Markets Act (DMA). The proposed measures aim to allow rival AI tools to interact more deeply with apps on Android devices, enabling tasks such as sending emails, sharing content, or placing orders while using a user’s preferred apps instead of relying mainly on Google’s own AI systems.

 

The Commission said this could give users greater choice by allowing competing AI services, including alternatives to Google’s Gemini, to integrate more easily into Android, including support for features such as custom wake words.

 

The EU is expected to finalise binding measures within six months, potentially reshaping how AI services operate across Android devices in the region.

 

It is possible that Apple may also be moving towards a more open AI ecosystem by allowing third-party AI models within Apple Intelligence.