Apple rolls out “Siri AI,” a new assistant built on Apple Intelligence and powered with help from Google’s Gemini

Apple used its WWDC 2026 keynote to unveil “Siri AI,” an entirely new version of Siri designed around Apple Intelligence and intended to be more conversational, more personal, and more capable across Apple’s apps and devices. In a June 8 press release, Apple said Siri AI includes a dedicated app for revisiting conversations across products, an expanded “Visual Intelligence” experience for understanding what is on screen, and integrated writing tools that help users draft and edit where they type. Apple also said the features are available for developer testing immediately and will reach users later this year as a beta, with early support limited to devices set to English.

Apple’s announcement arrives after years of pressure on the iPhone ecosystem to catch up with competitors’ generative AI assistants. Axios reported that Apple previewed its Apple Intelligence vision with the improved Siri at the center about two years ago, but struggled to deliver on promised capabilities on schedule, before later striking a deal with Google to use Gemini models to help power the effort. In its own materials for WWDC, Apple framed Siri AI as a “profoundly more capable” assistant that can draw on personal context understanding and systemwide capabilities, while maintaining what it describes as privacy protections built into the architecture.

What makes the WWDC reveal notable for the broader AI market is Apple’s explicit shift toward Google’s model ecosystem. In a joint statement published on Google’s blog, Apple and Google said they entered a multi-year collaboration in which next-generation Apple Foundation Models will be based on Google’s Gemini models and cloud technology, with the goal of powering future Apple Intelligence features including a more personalized Siri later this year. The company’s storyline at WWDC effectively presents Siri AI as the first major, user-facing expression of that partnership.

On iPhone, Apple said Siri’s multimodal capabilities are integrated into the Camera app through a new “Siri mode,” letting users tap to have Siri “see what they see” and respond with actions and information. Apple’s WWDC materials also described Visual Intelligence arriving beyond iPhone, with features integrated into iPad and Mac workflows, and extending to Apple Vision Pro so users can ask Siri about content they are viewing. For writing, Apple said Siri AI can generate and revise drafts by describing what the user wants, and can adapt tone and punctuation in messages and mail based on patterns from prior communications.

Apple’s rollout, however, is uneven across regions due to regulatory constraints. In a separate update also dated June 8, Apple said Siri AI will not be shipped to users in the European Union on iOS and iPadOS with the release of iOS 27 and iPadOS 27, citing the Digital Markets Act (DMA). Apple said it could make Siri AI available on macOS 27, visionOS 27, and watchOS 27, but claimed EU regulators did not accept its proposed solutions to allow the same capabilities for assistants while preserving privacy and security protections.

The push underscores how central the assistant layer is becoming to the iPhone value proposition, not just as a speech interface but as a control surface for other apps. By bundling conversation history in a dedicated Siri app, adding visual understanding, and expanding systemwide writing tools, Apple is positioning Siri AI as a way to reduce friction in everyday tasks, from finding information to updating content across apps. The company’s emphasis on using “on-device processing” and Private Cloud Compute in its EU delay explanation also indicates that privacy is a continuing part of its go-to-market narrative, even as it relies on third-party model technology for some foundation capabilities.

Still, several details remain unclear from Apple’s announcements. Apple did not specify what portion of Siri AI’s capabilities depend on Gemini-powered foundation models versus Apple-built components, nor did it provide concrete performance benchmarks, supported device lists by model, or information on compute and cost tradeoffs. In addition, while Apple described developer testing starting June 8 and a user beta “later this year,” it did not disclose the full schedule for broader language expansion beyond English, or what it will take to bring Siri AI to iPhone and iPad in the EU. What to watch next is whether developers can quickly integrate with Siri AI’s updated workflows, and how fast Apple can address region-specific regulatory hurdles while expanding availability.