Two big tech giants have joined hands: Apple and Google are collaborating to bring advanced AI to iPhones. Apple had previously announced it would use Google’s Gemini to power the next-generation Siri and other Apple Intelligence features, but the details were scarce. Now, a report by The Information reveals that Apple will have full access to the Gemini model in its own data centers, giving the company the ability to create smaller, task-specific models that can run directly on Apple devices.
Apple is gearing up for a major Siri overhaul, confirmed by Bloomberg’s Mark Gurman. The new features are expected to debut at WWDC 2026 on June 8, including a standalone app that gives Siri a fresh, chatbot-style experience.
Key features of the new Siri app:
Access all past conversations directly from the main interface.
Pin favorite chats and save older conversations.
Search across previous interactions.
Start new conversations via a plus (+) button.
Upload attachments such as images and documents.
Switch seamlessly between text and voice modes.
Receive suggested prompts based on prior usage.
Apple is also evolving Siri into a system-wide AI agent, enabling deeper integration across iPhones and Macs. The revamped assistant will:
Offer better control over device features and apps.
Access personal data such as messages, notes, and emails (with user permission).
Execute complete tasks within apps.
Access news and perform web searches using Apple’s built-in AI interfaces.
This redesign positions Siri as a more intelligent, proactive, and versatile assistant, combining chatbot capabilities with system-level functionality.
According to the report, Apple can use a process called distillation to transfer Gemini’s knowledge into smaller models. These compact models can run directly on devices, using less computing power, while mimicking Gemini’s “chain of thought.”
This also paves the way for a smarter Siri at WWDC 2026, capable of remembering past conversations and offering proactive suggestions. Meanwhile, Apple’s Foundation Models team continues to work on its own in-house AI models, though details on progress remain scarce.