Apple Intelligence Expands Across All Devices: What Changed in September 2025

Apple has rolled out Apple Intelligence features to iPhones, iPads, Macs, Apple Watches, and the Vision Pro headset. After initial announcements at WWDC25 in June 2025, these capabilities became widely available with Apple's September 2025 software releases. The rollout includes automation tools, real-time translation, photo editing, and home management features. Apple Intelligence is moving from a headline feature into something woven deeper into the operating system itself.

Shortcuts Gets Smarter

The most significant change is the addition of AI reasoning directly into Shortcuts — Apple's automation tool. Previously, users had to manually build step-by-step instructions to automate tasks. Now, you can describe what you want in plain English, and the system figures out which steps to take on its own. Apple calls these "intelligent actions."

Apple's September 2025 announcement confirms these intelligent actions rolled out with the September updates. The system uses AI models running on your device and on Apple's servers to understand your intent and then assembles the necessary actions to complete the task.

For app developers, this is a practical win. If they've already built their app to work with Apple's App Intents framework — a set of standards that lets apps declare what they can do — those capabilities become usable by the AI automation system automatically. Developers don't need to write extra code to support this. That makes it simpler for apps to participate in Apple's automation ecosystem.

Translation That Follows You Everywhere

Apple has expanded Live Translation across Messages, Phone, FaceTime, and AirPods on iOS devices. It translates both spoken and written language in real time, without sending your conversation to a third-party service. The AirPods addition is particularly interesting: it means your earbuds become the translation interface. In a face-to-face conversation, you can translate what someone is saying directly through your earbuds without pulling out your phone — a shift that makes sense for how people actually interact.

Apple hasn't revealed exactly how much translation work happens on your device versus on Apple's servers, though the company typically keeps latency-sensitive and privacy-sensitive tasks on-device. Full translation work probably uses a hybrid approach — some on your device and some on Apple's servers — but Apple hasn't confirmed those details publicly.

Cleaner Photos and Smarter Home Alerts

The Clean Up tool in Photos now uses AI to remove unwanted elements from pictures and fill in the background naturally. This is standard functionality in standalone photo apps, but having it built into Apple's stock Photos app normalizes the workflow for the hundreds of millions of people who never download a separate editing app.

In the Home app, Apple Intelligence now groups related notifications together and summarizes them. Instead of getting ten separate alerts from different sensors and devices, you get one summary that describes what actually happened in your home. This reduces clutter for people managing multiple smart home devices.

Who Can Actually Use This

Apple Intelligence requires Apple silicon — that means iPhone 15 Pro or later, or iPads and Macs with M-series chips. This hardware requirement matters when thinking about how many people can use these features. Apple's installed base of compatible devices is large, but it doesn't include every iPhone or Mac user. Some of the more demanding features may end up reserved for newer hardware over time.

The expansion to Apple Watch and Vision Pro broadens where Apple Intelligence appears, though the features available on those devices reflect what makes sense for their smaller screens, different interaction methods, and available computing power. A watch and a headset each open up different possibilities for AI assistance than an iPhone does.

What This Pattern Really Tells Us

Here's a broader perspective: over the past thirty years, I've watched Apple repeatedly absorb successful third-party features into its operating system. In the early 2000s, the dedicated media player category largely vanished once iTunes was built in. A decade later, flashlight apps, notification management, and QR readers went from independent apps to system features. The same gravity is pulling autonomous automation, translation, and generative image editing into the OS now.

This consolidation has real consequences. Developers who built standalone automation apps, translation services, or image editors face pressure from built-in alternatives. But the pattern also tends to improve what the broader ecosystem does. When baseline capabilities move into the OS, developers and companies compete on more ambitious, specialized use cases rather than reimplementing the same foundational tools.

For people who rely on automation and workflows, the natural-language Shortcuts capability is the biggest near-term change. Describing what you want instead of hand-coding steps removes a significant barrier to entry. However, there's a question worth flagging: Apple hasn't yet explained how transparent it will be about how the AI decides which steps to include in an automation. In every automated system built so far, users eventually want to see what decisions the AI made, why it made them, and how to fix mistakes. If that transparency doesn't appear, professional users may find the feature useful only for simple tasks, not the complex workflows they actually depend on.

The September 2025 releases mark a meaningful step toward Apple's goal of embedding generative AI throughout its platforms. The underlying architecture — AI running on your device where possible, Apple's Private Cloud Compute servers handling overflow, and a hardware baseline that ensures minimum computing power — stays consistent with what Apple outlined when it first introduced Apple Intelligence. What expands with each release is which parts of the system tap into that AI capability.