product updateApple

Apple Ships iOS 27 with Rebuilt Siri and 20+ New Apple Intelligence Features

TL;DR

Apple has released iOS 27 with a rebuilt Siri powered by large language models, plus more than 20 new Apple Intelligence features spanning Photos, Messages, Mail, Safari, and system-wide tools like proofreading and Call Context.

3 min read
0

Apple has released iOS 27, and according to Apple, the update includes a rebuilt version of Siri along with more than 20 new Apple Intelligence features spread across system apps and systemwide tools. The centerpiece is a ground-up overhaul of Siri, which Apple says now runs on large language model technology rather than its previous rule-based system.

Siri Gets an LLM Rebuild

Apple describes the new Siri as reconstructed "from the ground up for the LLM era." The company lists seven core capabilities:

  • Broad world knowledge — answers general questions beyond device-specific tasks
  • Personal context understanding — searches across Messages, Mail, Calendar, Notes, and Reminders to surface relevant information
  • Onscreen awareness — can reference and act on whatever content is currently displayed
  • App actions — executes tasks inside apps, such as sending messages or creating calendar events
  • Siri app — a dedicated app that syncs conversation history across devices via iCloud
  • Write with Siri — generates, edits, or gives feedback on text in virtually any text field
  • Expressive voices — adjustable pace and expressiveness on compatible hardware

Apple has not disclosed which model or model family powers the new Siri, nor has it published benchmark scores, parameter counts, or a training cutoff date for the underlying system.

App-Level Additions

Beyond Siri, Apple Intelligence features are now built into more first-party apps:

  • Photos: Spatial Reframing, Extend, improved Clean Up
  • Messages: Personalized Smart Reply
  • Calendar: natural-language event creation
  • Reminders: natural-language reminder creation
  • Mail: Suggestions and Personalized Smart Reply
  • Safari: Notify Me, natural-language extension description, tab organization by topic
  • Shortcuts: natural-language shortcut creation
  • Image Playground: redesigned interface, more realistic image generation, wallpaper creation
  • Files: intelligent file and folder naming suggestions
  • Home: combined activity notifications, AI-based search and captions for HomeKit Secure Video
  • Journal: customized writing prompts

Systemwide Features

Apple also added three features that apply across iOS rather than within a single app:

  • Automatic proofreading for spelling and grammar
  • Call Context, which surfaces relevant information onscreen during calls with businesses
  • Accessibility upgrades, including detailed image and scene descriptions in VoiceOver, flexible item naming in Voice Control, and content cleanup/summaries in Accessibility Reader

Apple has not specified hardware requirements for each feature individually, though the company notes that expressive Siri voices require "compatible hardware," suggesting device-tier gating similar to prior Apple Intelligence rollouts.

What This Means

This is Apple's most substantial Siri overhaul since the assistant's 2011 debut, and it arrives roughly two years after Apple first previewed Apple Intelligence and after well-documented delays to the LLM-based Siri rebuild. The breadth of the release — over 20 features across nearly every core app — signals Apple is now shipping AI as infrastructure rather than a bolted-on feature set.

What's missing is transparency that researchers and developers typically expect: no disclosed model architecture, no benchmark comparisons against competitors like Google Assistant or ChatGPT's voice mode, and no API access details for third-party developers. Apple's approach remains consumer-facing and closed, which makes independent verification of Siri's actual capabilities — versus Apple's marketing claims — difficult until reviewers and users test it against edge cases in personal context handling and onscreen awareness. For an assistant that has struggled with reliability in the past, execution quality will matter more than the feature count.

Related Articles

Comments

Loading...