Apple has released Siri AI, a rebuilt version of its virtual assistant powered by Apple Foundation Models developed in collaboration with Google's Gemini—a pairing that brings together two direct competitors in mobile operating systems. The rollout introduces what the company describes as the next generation of Apple Intelligence across iOS 27, iPadOS 27, macOS 27, watchOS 27, and visionOS 27, marking an architectural departure from previous iterations of Siri.

The redesigned assistant leverages deep personal context understanding to synthesize information across Messages, Mail, and Photos, enabling it to execute complex, multi-step queries that were previously beyond the scope of voice assistants. Apple's illustrative scenario has users ask Siri AI to identify a family activity mentioned in a text conversation, retrieve related instructions from a historical email, and automatically populate a grocery list in Reminders.

The system also introduces onscreen awareness, allowing it to interpret visible content and take contextual actions, alongside expanded systemwide app integrations that include third-party platforms such as WhatsApp and Audible, with Microsoft Outlook and Notability support forthcoming. Users can invoke the assistant through multiple entry points, including the Dynamic Island, Spotlight search, and systemwide context menus.

Siri AI begins rolling out today in English beta and will expand to French, Japanese, Korean, Portuguese, and Spanish next month, with no timeline given for languages beyond that initial set. Regulatory hurdles, however, have delayed availability in the European Union on iOS, iPadOS, and watchOS, as well as in China, where Apple cited ongoing compliance requirements—leaving users in two of Apple's largest markets without access to the rebuilt assistant at launch.

Multimodal Capabilities and Privacy-Centric Infrastructure

Alongside the new assistant, Apple is deploying Visual Intelligence across its hardware ecosystem, enabling multimodal queries through the iPhone Camera app, iPad screenshots, Mac keyboard shortcuts, and Vision Pro spatial computing. A standalone Siri app provides end-to-end encrypted, iCloud-synchronized conversation history so users can maintain contextual dialogue across devices, while enhanced Writing Tools integrate generative capabilities systemwide, including automatic proofreading and tone adaptation based on individual communication patterns.

The technical infrastructure relies on a hybrid processing model that combines on-device execution with Private Cloud Compute. Apple states that this extends the security architecture of the iPhone into cloud-based inference without exposing personal data to the company or third parties.

For the upcoming Apple Watch Series 12 and Ultra 4 hardware, Audio Intelligence features—including Live Rewind, for recalling recent conversation snippets, and Siri Recap, for generating summaries—will utilize a Secure Exclave on the S11 chip that processes audio in hardware isolation before immediately deleting it, with no persistent recordings created. That requirement ties the audio features to hardware that has yet to ship. The features operate on an explicit opt-in basis, give users granular control, and exclude sensitive identifiers from generated summaries.

Siri AI requires recent Apple silicon, including iPhone 16 or later, iPhone 15 Pro models, M1-equipped iP and Macs, and Apple Watch Series 9 or later—a hardware floor that excludes earlier iPhone, iPad, Mac, and Apple Watch models from the new assistant.

The integration of Google's Gemini models into Apple's proprietary stack, orchestrated through on-device systems such as Spotlight and App Toolbox, signals a notable evolution in consumer AI—one that attempts to reconcile advanced multimodal capabilities with hardware-enforced privacy guarantees and increasingly complex global regulatory constraints.

Source: Metaverse Post — Siri AI Debuts In English Beta Across Apple's 2027 Platforms, Excluding EU And China