Apple launches new Siri AI assistant with on-device processing and multimodal features, arriving in beta later this year

Apple announced a redesigned Siri powered chiefly by Google Gemini, with multimodal capabilities and integration across Apple devices. The new Siri will process user data on-device and via 'private cloud compute,' then discard it, according to executives. It launches in beta later this year, with no timeline for the EU and China.

The assistant can pull information from emails, texts, contacts, notes, calendars, and the internet to handle multi-step tasks—such as checking availability with a friend, setting calendar reminders, drafting messages that mimic your writing voice for specific recipients, or creating recipe lists and sending invitations. It includes on-screen awareness (the ability to understand what you're viewing), visual intelligence for image analysis, and Safari tab management.

Apple's AI strategy emphasizes privacy and integration into existing products rather than standalone capabilities. Unlike competitors courting enterprise users with expensive subscriptions, Apple is positioning AI as a pragmatic tool for everyday device owners—but the new features largely mirror functionality other companies have already introduced.