Apple's new Siri AI, powered by Gemini, finally arrives — but late and limited

Apple kicked off WWDC 2026 with a new Siri AI, its long-awaited answer to competitors like ChatGPT and Copilot. CEO Tim Cook promised technologies that “push the limits on what’s possible,” but the core reveal was a virtual assistant that finally catches up — years late. Siri is now backed by Apple’s own foundation models, but primarily powered by Google Gemini, a notable reliance on an outside partner. The company doubled down on privacy: agentic tasks run on-device or via “private cloud compute,” and user data is discarded afterward. On-screen awareness enables Siri to understand what’s on your display — for example, identifying a location in an Instagram photo and cross-referencing it with a friend’s address mentioned only in a text message. Other new abilities include drafting emails in your personal writing style (differentiating between messages to your boss vs. your best friend), adding calendar appointments, and suggesting smart replies. The Dynamic Island will now show AI-generated information cards tied to world events, weather, and your schedule. Despite these advances, Apple’s rollout remains cautious. The new Siri won’t be available until later this year, and only as a beta. No timeline exists for the EU or China, which Apple attributes to regulatory hurdles. Many of the announced features — like multi-step agentic workflows, on-screen awareness, and writing-style mimicry — have already been demonstrated by other companies. The question remains whether Apple’s years-late, privacy-first approach can win over users when the market has already moved on.