CtrlK
Apple Intelligence is Apple's personal intelligence system that integrates generative AI capabilities across iPhone, iPad, and Mac. Announced at Apple's Worldwide Developers Conference (WWDC) in June 2024, the system combines on-device machine learning models with cloud-based processing to deliver features including writing assistance, image generation, enhanced Siri capabilities, and intelligent notification management. Apple Intelligence is designed around a privacy-first architecture, processing most tasks directly on-device and routing more demanding workloads through a proprietary system called Private Cloud Compute.
Apple unveiled Apple Intelligence during its WWDC keynote on June 10, 2024, positioning it as "AI for the rest of us." The system was presented as a deeply integrated layer of intelligence woven into iOS, iPadOS, and macOS rather than a standalone application. Craig Federighi, Apple's Senior Vice President of Software Engineering, demonstrated the system's capabilities across writing, image creation, and an overhauled Siri experience during the presentation.[1]
The announcement came at a time when competitors like Google, Samsung, and Microsoft had already shipped consumer-facing AI features. Google's Gemini integration had been available on Pixel and Samsung devices for months, and Samsung's Galaxy AI suite had launched with the Galaxy S24 series in January 2024. Apple's approach differed by emphasizing privacy, on-device processing, and tight ecosystem integration over breadth of features.[2]
Apple Intelligence did not ship as a single release. Instead, Apple rolled out features incrementally across several software updates over the course of more than a year.
| Release | Date | Key Features Added |
|---|---|---|
| iOS 18.1 / macOS Sequoia 15.1 | October 28, 2024 | Writing Tools, new Siri design, Type to Siri, notification summaries, Smart Reply, Clean Up in Photos, call recording and transcription, Priority Messages in Mail |
| iOS 18.2 / macOS Sequoia 15.2 | December 11, 2024 | Image Playground, Genmoji, Image Wand, ChatGPT integration with Siri, Visual Intelligence (iPhone 16 only), Writing Tools enhancements |
| iOS 18.3 / macOS Sequoia 15.3 | January 2025 | Notification summaries disabled for News and Entertainment apps, new Visual Intelligence capabilities (plant/animal identification, calendar event creation from posters) |
| iOS 18.4 / macOS Sequoia 15.4 | March 31, 2025 | Expanded language support (Chinese, French, German, Italian, Japanese, Korean, Portuguese, Spanish, Vietnamese, and additional English locales), Priority Notifications improvements |
| iOS 26 (formerly iOS 19) / macOS Tahoe | September 2025 | Additional Apple Intelligence features, new foundation model updates, broader integration |
| iOS 26.4 / ongoing | Early 2026 | Continued feature rollout; some next-generation Siri features delayed to iOS 26.5 or iOS 27 |
The phased rollout drew criticism from consumers and media outlets who felt that Apple had marketed capabilities at WWDC 2024 that would not be available for many months, or in some cases, over a year after the initial announcement.[3]
Writing Tools is a systemwide feature available wherever text can be edited, including Mail, Messages, Notes, Pages, and third-party applications. It provides three core functions:
Writing Tools shipped with iOS 18.1 in October 2024 and received enhancements in iOS 18.2.[4]
Smart Reply is available in Mail and Messages. It analyzes incoming messages and suggests contextually appropriate responses that appear in the suggestions bar above the keyboard. In Mail, Smart Reply identifies questions within an email and ensures the suggested response addresses each one. The feature goes beyond generic quick replies by incorporating details about deadlines, availability, and other specifics referenced in the original message.[5]
Notification summaries use on-device language models to condense long or stacked notifications into brief, readable summaries displayed on the Lock Screen. This is particularly useful for active group chats or threads where dozens of messages may accumulate. The feature launched with iOS 18.1 but became controversial due to accuracy problems with news headlines (discussed in the Criticisms section below).[6]
Priority Messages appears as a dedicated section at the top of the Mail inbox, surfacing the most time-sensitive emails. The system identifies messages like same-day event invitations, boarding passes, or urgent requests and highlights them so users can act on them quickly without scrolling through less important mail.[5]
Genmoji allows users to create custom emoji by typing a description into the emoji keyboard. The system generates multiple options to choose from, and users can further personalize results by referencing photos of friends or family members from their photo library. The generated Genmoji can be used inline in messages, shared as stickers, or added as Tapback reactions. Genmoji shipped with iOS 18.2 in December 2024.[7]
Image Playground is a dedicated image generation tool available both as a standalone app and integrated into apps like Messages. Users create images by providing text descriptions and selecting from styles including Animation, Illustration, and Sketch. The app generates images locally on-device, and additional styles are available through ChatGPT integration. Image Playground was released with iOS 18.2.[7]
Image Wand is designed for the Notes app and works with Apple Pencil on iPad. Users can circle a rough sketch, and Image Wand transforms it into a polished image that matches the selected style (Animation, Illustration, or Sketch). Circling an empty space prompts Image Wand to generate an image based on the surrounding text and context in the note. It shipped alongside Image Playground in iOS 18.2.[7]
The Clean Up tool in the Photos app uses generative AI to remove unwanted objects and distractions from photographs. Users tap, brush over, or circle an element they want removed, and the system fills in the area with contextually appropriate content. After activating Clean Up, some objects are automatically highlighted for quick removal. The tool works best with smaller objects against uncluttered backgrounds. Clean Up was available from iOS 18.1.[8]
Visual Intelligence is exclusive to iPhone 16 models and uses the Camera Control button. Users long-press Camera Control to activate the feature, then point the camera at an object, sign, plant, animal, or scene. Visual Intelligence can summarize and copy visible text, translate text between languages, detect phone numbers and email addresses, create calendar events from posters and flyers, and identify plants and animals. The interface includes an "Ask" button that connects to ChatGPT for more complex questions and a "Search" button for Google image search. Visual Intelligence debuted in iOS 18.2 and received expanded capabilities in iOS 18.3.[9]
Starting with iOS 18.1, users can record phone calls directly within the Phone app. When recording begins, all participants are automatically notified. After the call ends, Apple Intelligence generates a written transcription and summary of the conversation.[4]
Apple Intelligence brought the most significant overhaul to Siri since its introduction in 2011.
Siri received a new visual design with a glowing light effect around the edges of the screen rather than the previous orb at the bottom. Users can also interact with Siri via typing ("Type to Siri") instead of voice, making it usable in situations where speaking aloud is not practical. These changes shipped with iOS 18.1.[4]
Apple announced that Siri would gain the ability to understand and act on content visible on the screen. For instance, a user viewing a restaurant in Safari could ask Siri to add the address to a contact card, or someone reading an article could ask Siri to summarize it. While this capability was demonstrated at WWDC 2024, the full implementation of on-screen awareness was delayed. As of early 2026, deeper on-screen awareness features remain in development and are expected to arrive in a future software update.[10]
Apple described a version of Siri that could draw on personal data, including messages, emails, calendar events, contacts, and photos, to answer questions like "What time is Mom's flight landing?" or "Pull up the PDF that Jamie shared last week." This personal context feature relies on on-device indexing and semantic search. Like on-screen awareness, the full rollout of personal context has faced delays, with Apple confirming that next-generation Siri features enabling deeper context awareness will arrive in 2026.[10]
Starting with iOS 18.2, Siri can optionally route queries to OpenAI's ChatGPT when a question falls outside Siri's built-in capabilities. Siri asks the user for permission before sending any information to ChatGPT. Users with a ChatGPT subscription can connect their account for access to paid-tier features. The integration allows Siri to tap into ChatGPT's broader world knowledge for complex questions, document analysis, and image understanding.[7]
Apple promised that Siri would be able to perform actions within apps, not just launch them. Examples include sending a specific photo to a contact via Messages, editing a document in Pages, or rearranging items in a to-do list. Third-party developers can expose app actions through Apple's App Intents framework. As with personal context, the full scope of in-app actions has been partially delayed.[10]
Apple Intelligence runs on a two-tier architecture: a compact on-device model for latency-sensitive tasks and a larger server-side model for more demanding workloads.
The on-device component is an approximately 3 billion parameter language model optimized for Apple silicon. It uses a transformer architecture with grouped-query attention and shared input/output embedding tables to minimize memory footprint. Two key innovations enable it to run efficiently on mobile hardware:
| Technique | Description | Impact |
|---|---|---|
| KV-Cache Sharing | The model is split into two blocks (62.5% and 37.5% of layers). Block 2 reuses key-value caches from Block 1. | Reduces KV-cache memory usage by 37.5% |
| 2-Bit Quantization-Aware Training | Model weights are compressed to a mixed 2-bit and 4-bit configuration, averaging 3.7 bits per weight. | Enables the 3B model to run on devices with 8 GB RAM |
On iPhone 15 Pro, the model achieves a time-to-first-token latency of roughly 0.6 milliseconds per prompt token and generates at approximately 30 tokens per second.[11]
Rather than maintaining separate fine-tuned models for each task, Apple uses LoRA (Low-Rank Adaptation) adapters. These are small neural network modules, typically tens of megabytes in size for the 3B model, that plug into the frozen base model's attention and feedforward layers. Different adapters handle different tasks (summarization, rewriting, code generation, etc.) and can be swapped dynamically at inference time. Adapter weights use 16-bit precision with a rank of 16.[11]
For tasks that exceed the on-device model's capacity, requests are routed to Apple's Private Cloud Compute (PCC) infrastructure. The server model uses a novel architecture called Parallel-Track Mixture-of-Experts (PT-MoE), which combines three techniques: track parallelism, mixture-of-experts sparse computation, and interleaved global-local attention. This design delivers high output quality at a competitive computational cost.
PCC runs on Apple-designed silicon servers. The system is built so that user data sent to PCC is used solely for the inference request, is never stored after the response is returned, and is not accessible to Apple employees, even those with administrative access. Independent security researchers can audit the software running on PCC nodes.[12]
Apple published technical details about its foundation models through its Machine Learning Research division. The models are trained on the AXLearn framework (built on JAX and XLA) using a combination of data parallelism, tensor parallelism, sequence parallelism, and Fully Sharded Data Parallel (FSDP). Training data includes licensed content, responsibly crawled web data collected via AppleBot, and high-quality synthetic data. Post-training refinement uses supervised fine-tuning and reinforcement learning from human feedback (RLHF) with mirror descent policy optimization.
In human evaluations, Apple's 3B on-device model outperformed several larger open-source models, including Phi-3-mini, Mistral-7B, Gemma-7B, and Llama 3-8B on summarization and instruction-following tasks. The server model performed comparably to GPT-3.5-Turbo and Llama 3-70B.[11]
Apple structured Apple Intelligence around three tiers of processing, each with distinct privacy properties:
| Tier | Processing Location | Privacy Guarantee |
|---|---|---|
| On-Device | User's iPhone, iPad, or Mac | Data never leaves the device. All processing uses the local 3B model and LoRA adapters. |
| Private Cloud Compute | Apple silicon servers in Apple data centers | Data is encrypted in transit, used only for the specific request, deleted immediately after response, and never accessible to Apple staff. Code is auditable by independent researchers. |
| Third-Party Models | External servers (e.g., OpenAI for ChatGPT) | Explicit user permission required before any data is sent. Apple states that third-party providers do not store requests. IP addresses are obscured. |
Apple emphasized that it does not use customer data to train its models and that the PCC architecture was designed to extend the security guarantees of on-device processing to cloud workloads. The company invited independent security researchers to verify these claims by inspecting the code running on PCC servers.[12]
Apple Intelligence requires hardware with sufficient memory and neural processing capabilities. All compatible devices have a minimum of 8 GB of RAM.
| Device Category | Compatible Models | Chip Requirement |
|---|---|---|
| iPhone | iPhone 15 Pro, iPhone 15 Pro Max, all iPhone 16 models, all iPhone 17 models (including iPhone 17e) | A17 Pro or newer |
| iPad | iPad Air (M1 and later), iPad Pro (M1 and later), iPad mini (A17 Pro and later) | M1 or A17 Pro or newer |
| Mac | All Macs with Apple silicon | M1 or newer |
The iPhone 15 and iPhone 15 Plus, despite running iOS 18, do not support Apple Intelligence because their A16 Bionic chip has only 6 GB of RAM and a less capable Neural Engine. The A17 Pro chip in the iPhone 15 Pro features a 16-core Neural Engine capable of approximately 35 trillion operations per second, which Apple identified as the minimum threshold for running the on-device model effectively.[13]
Apple's partnership with OpenAI brought ChatGPT integration to Apple Intelligence starting with iOS 18.2 in December 2024. The integration works in two primary ways: through Siri, which can forward complex queries to ChatGPT with user permission, and through Visual Intelligence on iPhone 16, where the "Ask" button sends captured images to ChatGPT for analysis. Users do not need a ChatGPT account to use the basic integration, though connecting a paid account unlocks additional capabilities.[7]
In January 2026, Apple announced a multiyear partnership with Google to integrate Gemini into Apple Intelligence. The deal, reported to be valued at approximately $1 billion per year, brings Google's Gemini models and cloud infrastructure into the Apple Intelligence stack. Gemini is expected to power an upgraded, more personalized version of Siri scheduled for release in 2026. Apple confirmed that the Google partnership does not replace the existing OpenAI relationship; both integrations will coexist, with Apple retaining the ability to route queries to the most appropriate backend model.[14]
The Google deal was widely interpreted as acknowledgment that Apple's own foundation models, while competitive for on-device tasks, needed supplementation from a leading cloud AI provider for the more ambitious Siri features Apple had promised at WWDC 2024.[15]
The most prominent controversy around Apple Intelligence involved its notification summaries generating inaccurate news headlines. In December 2024, a summary attributed to BBC News falsely stated that Luigi Mangione, the suspect arrested in the killing of UnitedHealthcare CEO Brian Thompson, had "shot himself." Another summary incorrectly claimed that darts player Luke Littler had won the PDC World Championship hours before the final was played. A third erroneously stated that tennis player Rafael Nadal had "come out as gay." The New York Times also experienced a false summary that read "Netanyahu arrested" when the actual story was about the International Criminal Court issuing an arrest warrant.[6]
The BBC issued a formal complaint, stating that the summaries "do not reflect, and in some cases completely contradict, the original BBC content." Reporters Without Borders expressed concern about "the risks posed to media outlets" by AI-generated summaries. In response, Apple disabled notification summaries for News and Entertainment apps in iOS 18.3 (January 2025) and stated it would work to improve accuracy before re-enabling them.[16]
Critics noted that many of the features showcased at WWDC 2024 were not available for months or, in some cases, over a year after the announcement. The more ambitious Siri capabilities (personal context, on-screen awareness, and in-app actions) were originally expected in 2025 but were delayed to 2026. As of March 2026, some of these features have still not shipped, with Apple reportedly pushing certain capabilities to iOS 26.5 or iOS 27.[3][10]
A law firm, CPM, opened an investigation into whether Apple's marketing of Apple Intelligence features on iPhone 16 models constituted misleading advertising, given that many highlighted capabilities were not functional at the time of the phone's sale.[3]
Apple Intelligence launched exclusively in U.S. English in October 2024. Localized support for additional English variants (UK, Australia, Canada, etc.) arrived in December 2024, but major languages including Chinese, French, German, Japanese, Korean, and Spanish were not supported until iOS 18.4 in March 2025. This left a large portion of Apple's global user base unable to use the features for roughly five months after launch.[17]
Apple Intelligence entered a competitive landscape where Google and Samsung had already established AI offerings on mobile devices.
| Feature Area | Apple Intelligence | Google Gemini / AI | Samsung Galaxy AI |
|---|---|---|---|
| Writing Assistance | Writing Tools (Rewrite, Proofread, Summarize) | Gemini in Google Docs, Gmail Smart Compose | Chat Assist, tone adjustment in Samsung Keyboard |
| Image Generation | Image Playground, Genmoji, Image Wand | Imagen integration in Google Photos and Messages | AI-generated wallpapers, stickers |
| Photo Editing | Clean Up (object removal) | Magic Eraser, Magic Editor in Google Photos | Photo Assist (move, resize, delete, and generate fill) |
| Visual Search | Visual Intelligence (iPhone 16 only) | Google Lens, Circle to Search | Circle to Search (Google-powered) |
| Translation | Live Translation in calls and messages (iOS 18.4+) | Google Translate integration, live caption translation | Live Translate for calls and text |
| Voice Assistant | Siri with ChatGPT/Gemini integration | Gemini as default assistant on Android | Bixby with Galaxy AI, Google Gemini available |
| Privacy Model | On-device processing, Private Cloud Compute | Cloud-based processing with Google's privacy policies | Cloud-based (primarily Google servers) |
Samsung's Galaxy AI, which launched with the Galaxy S24 in January 2024, gained an early lead in features like photo editing and real-time call translation. Google's Circle to Search became particularly popular, with surveys indicating that over 80% of Samsung users had tried the feature. Apple's competitive advantage centered on its privacy architecture and the depth of its ecosystem integration, though reviewers noted that Samsung produced superior results in photo editing tasks.[2][18]
As of March 2026, Apple Intelligence is available across all supported devices running iOS 26 (formerly iOS 19), iPadOS 26, and macOS Tahoe. The system supports over a dozen languages, with the March 2025 expansion to Chinese, French, German, Italian, Japanese, Korean, Portuguese, Spanish, and Vietnamese representing a significant milestone.
The Google Gemini partnership, announced in January 2026, is the most significant recent development. Gemini-powered Siri features, including deeper personal context awareness and cross-app task handling, are expected to debut later in 2026. Apple has confirmed these are coming but has not provided an exact release date, with reports suggesting some capabilities may slip to iOS 27 (expected September 2026).
Compatible devices now include the iPhone 17 series (including the budget-oriented iPhone 17e), all iPhone 16 models, iPhone 15 Pro and Pro Max, and all Macs and iPads with M-series or A17 Pro chips. The ChatGPT integration remains operational alongside the new Google Gemini partnership.
Notification summaries for news apps remain a work in progress, with Apple continuing to refine the feature's accuracy before fully re-enabling it across all app categories. The broader suite of Apple Intelligence features, including Writing Tools, Image Playground, Genmoji, Clean Up, and Smart Reply, is stable and widely used.[19][20]