Apple launches new Siri based on Gemini but excludes European market

This week, Apple released the beta version of its revamped virtual assistant, which uses Google's Gemini models to process requests. According to The Decoder, the new Siri can now analyze on-screen content and understand personal context extracted from messages, emails, and photos, allowing for task execution across different applications. The initial launch is limited to English, with support for Portuguese, Japanese, Korean, French, and Spanish expected next month.

Technological integration and system operation

The new Siri operates on a hybrid architecture, splitting data processing between the local device and Private Cloud Compute. According to The Decoder, Apple guarantees that personal data is neither stored nor made accessible to the company. The tool allows for interactions within third-party apps such as WhatsApp and Audible, with expansion planned for Outlook, Notability, and Tripsy. In addition to voice commands, the assistant can be activated via the side button, the Dynamic Island, or Spotlight on iPad and Mac.

Geographic limitations and initial feedback

Not all users will have immediate access to these new features. The Decoder reports that Apple has excluded the European Union and China from the initial launch, citing privacy and security concerns in the former and pending regulatory requirements in the latter. Early user tests have yielded mixed signals. While some report practical gains in multi-step tasks, such as organizing family recipes, others, like MacStories, point to failures in interpreting simple personal contexts that should already be accessible to the assistant. The Gemini model, which replaces the foundation of the old Siri, still exhibits hallucinations and misunderstandings, a reminder that the technology, while advanced, still faces accuracy challenges that persist in daily use.