Apple's Siri Upgrade May Compromise Privacy with Google Cloud

Apple’s long-delayed Siri overhaul, aimed at catching up to modern AI chatbots, may come with significant privacy compromises, according to a Bloomberg report by Mark Gurman. Apple quietly partnered with Google in January to use its AI technology for Apple Foundation Models, and now sources indicate that parts of the new Siri will actually run on Google's cloud infrastructure. This is a departure from Apple's longstanding practice of processing sensitive data like face recognition and health information directly on-device with strong encryption. While Apple still touts its Private Cloud Compute system for AI tasks, the company has not clarified whether Google's chips and data centers will handle security protections for Siri data. That ambiguity has privacy advocates worried, especially as Siri gains memory features that let it recall conversation context — a capability that typically requires storing user thoughts and habits in the cloud.

The new Siri will reportedly allow users to choose how long their conversation data is retained: 30 days, one year, or forever. But such options may not fully address fears that intimate user queries could end up on “a Google server somewhere.” Gurman notes that Apple plans to emphasize privacy at this year’s Worldwide Developers Conference starting June 8, arguing its approach differs from rivals that broadly train models on user interactions. Still, the company’s reliance on Google’s infrastructure — rather than its own — marks a notable erosion of its historic privacy-first stance. For years, Apple has marketed itself as the pro-privacy alternative in big tech; this shift could alienate loyal users who expect every aspect of Siri to stay within Apple’s walled garden.