Apple Intelligence is a suite of AI-powered features available on iPhone, iPad and Mac. The platform utilizes on-device computation as well as server-based processing.
Apple Intelligence enables system-wide writing tools, summarization, visual look-up, image generation, automation and more. An additional layer of functionality is integration with OpenAI’s ChatGPT.
Apple Intelligence was announced during the WWDC24 keynote, but was not present in the first version of iOS 18 in September. The first Apple Intelligence features, like Writing Tools and Summaries, are available in iOS 18.1, iPadOS 18.1 and macOS 15.1 Sequoia. Updates in iOS 18.2 and macOS 15.2 add even more features, like Genmoji and Image Playground. visionOS 2.4 will bring Apple Intelligence to the Vision Pro.
Table of Contents:
To preserve user privacy, Apple Intelligence works in three key ways. Apple says its unique method will produce highly personal results while maintaining the highest level of data privacy.
Apple Intelligence is powered by two foundation models, a large-language model and an image diffusion model. A system of light-weight adapters can plug into these models to power all kinds of different features.
Two versions of the Apple Foundation model (AFM) were created to run on-device and in the cloud.
The first level of AI computation takes place on the Apple device itself. For that reason, Apple Intelligence requires recent hardware. Only Macs and iPads with Apple’s M-series chips, along with the iPhone 15 Pro and iPhone 16 lineup, can run Apple Intelligence.
A semantic index can tap into the data that third-party apps store on your device. This can include records of your email, conversations, files, photos, contacts and more. The semantic index will be used to power the knowledge Siri can build about you to answer more useful questions about your life, like an all-knowing assistant.
Also stored on-device is the app intents toolbox. This is a list of features and capabilities that are offered by the apps you have installed on your device. Siri will be able to control apps using features in the toolbox, expanding the voice assistant to handle many more kinds of tasks.
For tasks requiring more computing power, Apple created Private Cloud Compute. This sends the user’s query to Apple servers, with end-to-end encryption to preserve privacy.
Apple designed this off-device computing platform without persistent storage. No user information stays on Apple servers, minimizing the possibility of a data breach. Sebastien Marineau-Mes, Apple’s vice president of intelligent system experience engineering, says this unprecedented level of security in cloud computing makes it “the most advanced security architecture” for a cloud service like this.
The software images that power Private Cloud Compute servers are publicly available. This allows security researchers to analyze and verify Apple’s security claims.
ChatGPT is integrated into two Apple Intelligence features, starting with iOS 18.2. When asking Siri a general knowledge question (outside Siri’s current integration with Wikipedia, math and movie, TV and sports trivia), you can ask ChatGPT. And for text generation, ChatGPT is integrated into the system writing tools.
Each query requires explicit permission to pass the user’s info to OpenAI’s chatbot.
Anyone can use these features for free, without creating an account. If you have a ChatGPT account and want to access paid features and higher usage limits, you can sign in.
The features will be powered by OpenAI’s impressive GPT-4o. While terms of the controversial partnership between Apple and OpenAI remain confidential, no money has reportedly changed hands.
In iOS 18.1, Siri has a few new conveniences:
iOS 18.2 adds integration with ChatGPT:
A more substantial update, likely arriving early next year, will provide a more substantial revamp of versatile new capabilities:
A future update in iOS 19, likely arriving in spring 2026, will totally overhaul Siri with a more conversational dialogue.
Visual intelligence lets you point your camera at something, press the Camera Control button, and look it up:
Visual Intelligence available on the iPhone 16 and iPhone 16 Pro models activated by the Camera Control. On the iPhone 16e and iPhone 15 Pro, it will be available via the Action button.
The AI-powered Writing Tools in Apple Intelligence function system-wide, including in third-party apps. They include: Rewrite, Proofread and Summarize. Similar to competitors’ AI-powered writing tools, they will give users the ability to quickly improve their written words.
When replying to text message or email, you’ll see suggested replies. It can also make sure your reply addresses every question you’ve been asked.
Image Clean Up is a new way to edit your pictures. If you have a picture that’s otherwise great, aside from one distracting thing in the background, it’s easy to remove. Just circle part of an image to have AI fill it in. It’s a new tool that you can access just by tapping the Edit button in Photos, alongside cropping, red eye and other image tools.
If you’re trying to find a specific photo or video, you can now search using natural language, like “video of Kim skateboarding wearing a blue sweater.”
You can create memory movies with a natural language prompt. Apple Intelligence understands the context of your request, the type of images it should look for, the people in your photos, the story told through the sequence of image, and even the right Apple Music track to set it to.
Apple Intelligence can generate AI images on-device from Messages, Notes and Apple’s new Image Playground app. Users enter a prompt describing the image they want, and Image Playground creates images in one of two styles: Animation and Illustration.
A similar feature, Image Wand, will let you turn your sketches or pages of notes into a drawing using a third style, Sketch.
Apple Intelligence allows users to create custom emoji called Genmoji. The highly personal emoji, created via simple prompts, can be based on photos of the user or the user’s contacts. And they can be shared using Apple’s Messages app (including as custom Tapback reactions) as well as updated third-party apps. You can create them just by typing descriptions. Users can pick people in photos and create Genmoji images that resemble them.
If you get a big string of notifications in a row, they’ll be condensed into a summary. It can give you the gist of a particularly noisy group chat without reading through every message.
If a bunch of notifications build up, Priority Notifications will make sure the important ones don’t get buried.
A new Reduce Interruptions Focus mode will intelligently choose which notifications to let through.
In Mail, instead of seeing a preview of the first two lines of an email, you’ll see a one-sentence summary of the contents. Emails for two-step verification codes, for example, will always show you the code right in the notification banner.
Mail will also have your most important emails put at the top of your inbox.
For software developers, Apple Intelligence brings AI-powered code completion to Xcode, the company’s integrated development environment for Mac. Apple trained Xcode 16’s AI model on the Swift coding language as well as the company’s APIs. It should simplify app development by using AI to properly complete code.
A larger and more powerful cloud-based model, called Swift Assist, will enable fast prototyping of apps. It can generate code as well as sample data that devs can use as they create apps.
No Apple Intelligence features are present in iOS 18.0.
iOS 18.1, released on October 28, 2024, introduced the following features:
iOS 18.2, released on December 11, 2024, adds more features:
iOS 18.3 brings dog, plant and event identification to visual intelligence and enables Apple Intelligence by default.
The following features will be added in a future update to iOS:
In the further future, Siri may have a total LLM-powered overhaul to be more conversational.
As many Apple Intelligence features run locally on-device, only the latest devices will support it. Apple Intelligence only runs on Apple silicon chips with a recent Neural Engine and 8 GB of unified memory.
Apple Intelligence runs on iPhones with an A17 Pro chip or later.
Apple Intelligence runs on iPads with an M-series or A17 Pro chip.
The entry-level iPad does not run Apple Intelligence, as it uses an older A16 chip and doesn’t have enough memory.
Apple Intelligence runs on any Mac with Apple silicon.
Apple Intelligence will come to visionOS 2.4 in April, 2025.
In iOS 18.1, Apple Intelligence is only available in U.S. English. Support for other regions will roll out over time.
Apple Intelligence will become available in English localized to Australia, Canada, New Zealand, South Africa and the United Kingdom in iOS 18.2.
In April 2025 with iOS 18.4, Apple will expand this list further by supporting Chinese, English (India), English (Singapore), French, German, Italian, Korean, Japanese, Korean, Spanish and Vietnamese languages.
In European Union member countries, Apple Intelligence is only enabled on the Mac. Users can circumvent this by setting their iPhone or iPad’s region to any non-EU member country and their language to any of those supported. Support will officially come with iOS 18.4.
With visionOS 2.4, Apple Intelligence will run on the Vision Pro in U.S. English.
Apple continues beta testing iOS 26.1, dropping the third beta to developers. It builds on the changes in the previous two betas, only introducing minor new changes.
Below is everything new we have found in iOS 26.1 beta 3 so far.
Apple is already working on future generations of its AirPods lineup despite only recently releasing the AirPods Pro 3, according to a new report. The company focuses development efforts on transforming its popular earbuds into AI-powered devices with enhanced health tracking capabilities.
Apple is reportedly testing an AI-powered chatbot that can perform many tasks that the promised Siri upgrade will be able to handle. The app supposedly gives Apple employees a way to test future Siri capabilities, but also to explore whether Siri needs a chatbot functionality.
ChatGPT, Google Gemini and others made AI chatbots mainstream, so not having one in the long-promised Siri revamp would be a glaring omission. One that would draw considerable criticism.
Apple’s new feature that lets AirPods act as language translators isn’t available for EU residents while they’re in Europe. What makes the restriction odd is that the feature can be used by EU residents outside of the European Union.
And Live Translation with AirPods can be used by those traveling to the European Union from the United States, the United Kingdom, etc.
You don’t have to buy the just-announced AirPods Pro 3 just to get Apple’s new Live Translation feature. The AI-powered communication option is coming to older models of wireless earbuds, too.
An iPhone that supports Apple Intelligence is required as well.
Apple will reportedly take on OpenAI and Perplexity with its own AI-powered web search. Known internally as the World Knowledge Answers, the feature should launch by next year.
The company will integrate it directly into Siri and may eventually bring it to Safari and Spotlight.
You can combine two or more emoji into your own custom creation in iOS 26 using Apple’s upgraded Genmoji tool. In my testing, it’s great at adding sunglasses and cowboy hats to other emoji. It’s also pretty good at converting ordinary yellow-face emoji into cats, frogs, skulls, etc.
You can easily make a pumpkin cowboy, keyboard cat, pregnant Santa, banana phone or Easter Island statue snowman … the possibilities are endless. Well, not entirely endless — as fun as it would be to play geopolitics by combining national flags, all the flag emoji are restricted.
The best Apple Intelligence feature gets even more fun in iOS 26. Here’s how to use Genmoji‘s fun new feature.
Google’s new Pixel 10 lineup shows just how much fun (and just how useful) AI can be on a phone. If Apple really wants the iPhone 17 lineup to shine, it should steal a few pages from Google’s playbook.
The iPhone 17 needs more than impressive new hardware to stand out from the crowd. It needs new AI features to make an impact, especially with the Pixel 10 raising the bar.
Add Google Gemini to the list of AIs that might go into the promised revamp of the Siri voice assistant if Apple can’t develop the tech itself.
Apple is working hard on its own large language models to provide Siri with a much-needed intelligence boost. But Cupertino is reportedly covering its bases by talking to other companies about using their AI instead. Google joined the list, according to an unconfirmed report published Friday.
Apple’s new internal team, “Answers, Knowledge and Information,” is apparently developing a ChatGPT-like search tool. It will reportedly use an “answer engine’ that will crawl the web to answer simple search querries.
The tool is still in the early stages of development, so it won’t go live anytime soon.
In an unusual all-hands meeting Friday, Apple CEO Tim Cook assured employees that Apple won’t drop the ball when it comes to artificial intelligence. Calling AI “as big or bigger” than the internet, Cook said the company will rise to the occasion.
“Apple must do this,” he said. “Apple will do this. This is sort of ours to grab. We will make the investment to do it.”
In addition to hyping the company’s AI efforts, Cook expressed excitement about all the “amazing” new Apple products in the pipeline. And Apple software chief Craig Federighi told his colleagues not to worry about the long-delayed smarter Siri — a key component of Apple’s AI-infused future.
Apple has an AI problem, and it’s hoping to solve it in the traditional manner: by pouring money on it.
In a conference call with investors Thursday, Apple CEO Tim Cook acknowledged that the company’s costs for developing artificial intelligence technology continue to grow. He also mentioned that the company is not just open to acquiring companies with AI expertise — it’s already actively doing so.
Nevertheless, the AI-enhanced version of Siri won’t arrive before 2026.
Apple faces big challenges in its artificial intelligence efforts as another key researcher leaves the company to join Meta’s ambitious superintelligence project, according to a new report Tuesday. The departure marks the fourth AI expert to leave Apple’s foundation models team in just one month, raising questions about the future of Apple Intelligence and the company’s AI strategy.
Apple’s newest retail destination opens its doors to customers Saturday in the heart of Osaka, Japan. That’s two decades after its first store opened in the city, located in Japan’s Kansai region. The new Apple Umeda store promises to deliver the full Apple experience to one of the country’s most vibrant commercial districts. It includes an exclusive Vision Pro demo area and a Today at Apple session on Apple Intelligence.
“I am very pleased to open Apple Umeda, a wonderful space that reflects the energy of the vibrant city of Osaka,” said Deirdre O’Brien, Apple’s Senior Vice President of Retail and Human Resources. “This new store is a place to walk with creative communities and local businesses in Kansai.”
Apple’s AI ambitions take a hit as a key executive leading its AI model efforts departs the company. Meta has successfully managed to poach Ruoming Pang, who managed Apple’s foundation models team.
Meta is forming a new superintelligence team and reportedly offering pay packages of up to $100 million to attract top AI talent.
Apple acquired the brilliant image-editing app Pixelmator in November 2024. And now, eight months later, it’s updating the app with Apple Intelligence support.
For now, AI integration remains exclusive to Pixelmator Pro on Mac.
Apple’s struggle to develop artificial intelligence might reach the point where the iPhone-maker will need to outsource one of its core technologies: Siri. A promised AI upgrade for the voice assistant may be powered by large language models created by Anthropic or OpenAI, not Apple itself, according to an unconfirmed report published Monday.
But this is only a possibility — no decision has been made.
Apple reportedly held internal discussions to acquire AI startup Perplexity AI. The latter is an AI-powered search engine. It uses a large language model (LLM) to process the answers and presents them in an easier-to-understand format.
The discussions inside Apple are seemingly at an early level, and it may not even officially provide an offer to the young startup.
iOS 26, iPadOS 26, watchOS 26 and macOS 26 introduce a host of new features in addition to the Liquid Glass design makeover that will bring glossy, translucent harmony to Apple’s software ecosystem this fall. However, not all of those advanced features will make it to every compatible device when Apple releases its next-gen operating systems.
If you own an old iPhone or Intel-based Mac, you will miss out on several improvements. Here’s a look at what won’t make the cut.
Despite all the criticism, mockery and predictions of doom, the fact that Apple is lagging the pack in artificial intelligence isn’t a catastrophe. The AI boom has barely started, and average consumers remain doubtful about the technology.
More importantly, an Apple executive points out that the company doesn’t need to develop its own cutting-edge AI to benefit from the research done by other companies.
Craig Federighi, Apple’s head of software development, finally gave more details on why the AI-enhanced new Siri version got delayed for so very long. He explains why Apple thought it could deliver the new feature this year, but eventually had to push back the launch until possibly 2026.
The intent seems to be to show that Apple was being overly optimistic, not deceptive, when it said at WWDC24 the new Siri version would be out within a year.
The highlight of Apple’s recent AI efforts is Live Translation, but that’s not the only new Apple Intelligence feature unveiled at WWDC on Monday. There are also improvements to visual intelligence and Image Playground. Plus, third-party app developers can access Apple’s AI models for free.
But the keynote address kicking off the Mac maker’s developers conference was short on big AI-related announcements when compared to what’s coming out of OpenAI or Google. Still, the company did what it could.
“Last year, we took the first steps on a journey to bring users intelligence that’s helpful, relevant, easy to use, and right where users need it, all while protecting their privacy. Now, the models that power Apple Intelligence are becoming more capable and efficient, and we’re integrating features in even more places across each of our operating systems,” said Craig Federighi, Apple’s senior vice president of Software Engineering.
Apple plans to devote a considerable portion of WWDC25 to talking about its AI strategy, according to a reliable source. That’s something of a surprise, as the iPhone maker has no big AI-related announcements to make at the developer conference.
Siri and Apple Intelligence will even get a rebrand, according to this source.
Apple’s Image Playground is a free app for iPhone and other Apple devices that lets you generate unlimited AI images for free. You can generate images based on your friends, from a text prompt or totally from scratch, in a wide variety of themes and styles.
It’s part of Apple Intelligence, the growing set of AI features that work on the latest iPhones, Macs and iPads. Here’s how to use Image Playground.
Apple’s much-anticipated launch of Apple Intelligence in China has been indefinitely delayed as the company’s AI partnership with Alibaba faces regulatory roadblocks tied to the intensifying trade war between the United States and China, according to a new report.