This approach aligns with Apple’s long-standing privacy strategy, contrasting with competitors like Microsoft, whose Copilot+ standard also supports on-device AI processing but focuses on laptops
Apple has announced its next big leap forward: Apple Intelligence. This new feature will integrate artificial intelligence into your iPhone, Mac, or iPad, promising a smarter Siri, improved email and essay writing, unique emoji creation, and the transformation of rough sketches into polished AI art. Set to launch later this year, Apple Intelligence aims to be one of the most user-friendly and intuitive generative AI implementations yet.
However, there’s a catch. Not all Apple devices will support Apple Intelligence. Only users with an iPhone 15 Pro or iPhone 15 Pro Max can access the new features. Regular iPhone 15 users are left out, making their devices feel outdated in less than a year. Mac users, on the other hand, need an Apple Silicon computer, which includes models released since 2020.
Some speculate that these restrictions are tied to typical upgrade cycles. While it’s common for people to upgrade their phones annually, doing the same with laptops is less frequent and often seen as impractical. But there’s more to it. The computational power required for Apple Intelligence is different from typical tasks performed on iPhones or Macs.
When using services like ChatGPT or Adobe Photoshop’s Generative Fill, the heavy lifting is done by remote cloud servers, not the device itself. Apple Intelligence aims to change this by performing AI tasks directly on the device.
Apple emphasises privacy with this new feature. “You should not have to hand over all the details of your life to be warehoused and analyzed in someone’s AI cloud,” said Craig Federighi, Apple’s senior vice president of software engineering. “The cornerstone of the personal intelligence system is on-device processing.”
This approach aligns with Apple’s long-standing privacy strategy, contrasting with competitors like Microsoft, whose Copilot+ standard also supports on-device AI processing but focuses on laptops.
Central to this development is the Neural Processing Unit (NPU). NPUs handle numerous operations simultaneously while using minimal power, essential for the AI tasks in question. They excel at matrix manipulation, the core of how AI models generate content.
Apple introduced its first NPU, the Apple Neural Engine, in 2017 with the iPhone 8 and iPhone X. Since then, iOS has utilised AI and machine learning extensively. But Apple Intelligence requires more robust hardware, particularly more RAM. Older iPhones typically have 6 GB of RAM, whereas devices compatible with Apple Intelligence have 8 GB. This additional memory is crucial for running AI models locally, as these models need to be stored in RAM for efficient processing.
Despite the advances, there are still limitations. Even high-end devices like the iPhone 15 Pro Max have only 2 to 4 GB of RAM available for AI tasks and a 35 TOPS-rated NPU. This suggests that more complex AI functions will still rely on cloud processing.
Apple has stated that certain features, like Sharing Suggestions and scene recognition in Photos, will be processed on-device. However, the more advanced generative AI features likely cannot be handled entirely locally on an iPhone.
The real test will come with the release of iOS 18 later this year, coinciding with the launch of the iPhone 16. It remains to be seen whether the new iPhones will support the full suite of Apple Intelligence features or if Apple will continue to reserve some capabilities for its Pro models.
As the tech world anticipates this rollout, users will soon experience firsthand how Apple’s latest AI advancements impact their daily interactions with their devices.