SourceStalecollected in 1m

Apple’s New Siri App to Feature Auto-Deleting Chats

Read original on Bloomberg Technology
#privacy#ios#generative-ai

Apple's strategy for balancing generative AI with privacy-first data retention is critical for enterprise adoption.

30-Second TL;DR

What Changed

New Siri application will feature ChatGPT-like conversational capabilities.

Why It Matters

This move signals Apple's commitment to integrating generative AI while maintaining its strict stance on data privacy. It sets a standard for how consumer-facing AI assistants should handle ephemeral data.

What To Do Next

Monitor Apple's developer documentation for new SiriKit APIs that support ephemeral data handling for LLM-based features.

Who should care:Developers & AI Engineers

Key Points

  • New Siri application will feature ChatGPT-like conversational capabilities.
  • Privacy-focused auto-deleting chat history will be implemented.
  • Genmoji functionality is scheduled for an upgrade in iOS 27.

Deep Insight

Background and context from public sources — not the original article. 12 sources cited.

Enhanced Key Takeaways

  • The new Siri is integrated into a broader system called "Apple Intelligence," which leverages generative AI across iOS, iPadOS, and macOS, utilizing both on-device processing and a secure "Private Cloud Compute" for complex tasks to ensure user privacy.
  • Apple has partnered with Google to incorporate its Gemini AI model to power some of the new Siri features, aiming to combine Google's advanced AI capabilities with Apple's stringent privacy protections.
  • Genmoji, a feature introduced as part of Apple Intelligence, enables users to create personalized emojis by describing them with text prompts, combining existing emojis, or generating them based on personal photos.
  • The enhanced Siri will feature "on-screen awareness" and "personal context" across various Apple applications like Mail, Calendar, Photos, and Notes, allowing it to perform multi-step, cross-app actions.
  • Apple's privacy framework for Siri includes associating request history with a random, device-generated identifier that frequently rotates, rather than linking it to an Apple Account, and does not retain audio recordings by default unless explicitly opted in.

Competitor Analysis

Core Capability
Apple Siri (with Apple Intelligence)
Conversational AI, generative AI, on-device/cloud processing, cross-app actions, Genmoji
OpenAI ChatGPT
Conversational AI, text generation, code generation, image generation (DALL-E)
Google Assistant (with Gemini)
Conversational AI, proactive assistance, deep Google ecosystem integration, task automation
Amazon Alexa
Conversational AI, smart home control, shopping, entertainment
Privacy Model
Apple Siri (with Apple Intelligence)
On-device processing, Private Cloud Compute for complex tasks (data not stored), random identifiers for Siri requests, auto-deleting chat history
OpenAI ChatGPT
Data used for model training by default (opt-out available), chat history stored
Google Assistant (with Gemini)
Collects voice data associated with individual accounts, opt-out for recordings available
Amazon Alexa
Collects voice data associated with individual accounts, opt-out for recordings available
Ecosystem Integration
Apple Siri (with Apple Intelligence)
Deep integration across iOS, iPadOS, macOS, Apple apps (Mail, Calendar, Photos, Notes)
OpenAI ChatGPT
Web-based, API access, some third-party integrations
Google Assistant (with Gemini)
Deep integration across Android, Google apps (Gmail, Calendar, Drive, Photos, YouTube)
Amazon Alexa
Amazon ecosystem (Echo devices, shopping, music)
Customization/Extensibility
Apple Siri (with Apple Intelligence)
"AI Extensions" for third-party models in iOS 27 expected, Genmoji creation
OpenAI ChatGPT
Plugins, custom GPTs
Google Assistant (with Gemini)
Custom homescreen widgets, Wear OS Tiles (Android)
Amazon Alexa
Skills
Pricing
Apple Siri (with Apple Intelligence)
Free with supported Apple devices
OpenAI ChatGPT
Free tier, ChatGPT Plus subscription ($20/month)
Google Assistant (with Gemini)
Free with Google accounts/devices
Amazon Alexa
Free with Amazon accounts/devices
Benchmarks
Apple Siri (with Apple Intelligence)
Not publicly disclosed for direct comparison
OpenAI ChatGPT
Various LLM benchmarks (e.g., MMLU, GSM8K)
Google Assistant (with Gemini)
Various LLM benchmarks (e.g., MMLU, GSM8K)
Amazon Alexa
Not publicly disclosed for direct comparison

Technical Deep Dive

  • Apple Intelligence, which powers the new Siri, utilizes a combination of on-device processing and server-based computation.
  • It leverages Apple silicon to efficiently handle the understanding and creation of language and images.
  • The core large language model within Apple Intelligence is accessible to developers through the Foundation Models framework, enabling on-device privacy for app-specific intelligent features.
  • For complex requests, Apple employs "Private Cloud Compute," which extends the privacy and security of Apple devices into the cloud, ensuring that user data is never stored and is used solely to fulfill the immediate request.
  • Siri's request history, including transcripts and related data, is associated with a random, device-generated identifier that rotates multiple times per hour and is not linked to a user's Apple Account or email address.
  • By default, Apple does not retain audio recordings of Siri and Dictation interactions; users must opt-in for Apple to store and review audio, and even then, these samples are associated with a random identifier.
  • Apple has confirmed a partnership with Google to integrate Google's Gemini AI model to power certain new Siri features.

Future ImplicationsAI analysis grounded in cited sources

Apple's emphasis on on-device processing and Private Cloud Compute will set a new industry standard for AI privacy.
By processing many requests on-device and using a secure, non-storing cloud for complex tasks, Apple aims to differentiate its AI offerings through superior user data protection, potentially pressuring competitors to adopt similar privacy-centric architectures.
The integration of Google's Gemini AI with Apple's ecosystem will significantly accelerate Siri's capabilities and user adoption.
Leveraging an established, powerful LLM like Gemini allows Apple to quickly enhance Siri's conversational and contextual understanding without solely relying on its own nascent models, potentially leading to a more robust and competitive assistant.
The anticipated "AI Extensions" feature in iOS 27, allowing users to choose third-party AI models, will foster a more open and competitive AI ecosystem on Apple devices.
Giving users the choice to select AI models beyond Apple's default or ChatGPT for various tasks could encourage innovation from other AI developers and provide users with greater flexibility and customization.

Timeline

2005
Development of Siri by SRI International AI Center begins.
2010-04
Apple acquires Siri Inc., the company behind the standalone Siri app.
2011-10
Siri debuts as a built-in feature with the iPhone 4S.
2019-08
Apple revises Siri privacy policies, allowing users to opt out of audio recording reviews after public concerns.
2024-06-10
Apple introduces Apple Intelligence, a generative AI system deeply integrated into iOS, iPadOS, and macOS, featuring an overhauled Siri and Genmoji.
2026-01
Apple confirms partnership with Google to utilize Gemini AI models for new Siri features.

Event Coverage

Weekly AI Recap

Read this week's curated digest of top AI events →

AI-curated news aggregator. All content rights belong to original publishers.
Original source: Bloomberg Technology

This is a summary, not the original. Read the source, or get the weekly briefing.

The weekly digest

One email a week. Unsubscribe anytime.