Apple made a big announcement this week about AI features coming to its operating systems this fall. But most users won’t have access to them.
Apple’s big announcement at the Worldwide Developers Conference this week was the presentation of a suite of features that they are calling “Apple Intelligence.” These artificial intelligence (AI) features will help people write, create images, edit photos, and will give superpowers to Siri. But most users won’t be able to get these features this year or in the immediate future.
To begin with, these features are limited to US English at launch, though people in other English-speaking countries will have the option to switch the language of their devices to try them out. Adding other languages will take time.
The bigger issue is that not many Apple devices will be compatible with these features. As Apple says, “Apple Intelligence will be available in beta on iPhone 15 Pro, iPhone 15 Pro Max, and iPad and Mac with M1 and later.” This means that, aside from Macs and iPads, only one iPhone model—or the two variants of one model—will be able to run Apple Intelligence.
In February, Apple said that there are 2.2 billion active Apple devices around the world. Estimates suggest that there are about 1.46 billion active iPhone users in the world, and, of these users, a small number have iPhone models that will run Apple Intelligence. The iPhone 15 Pro and iPhone 15 Pro Max are estimated to make up about 45% of total iPhone sales, and, while Apple doesn’t report the number of units the company sells, estimates are around 230 million per year.
If you do some rough math, one could speculate that perhaps 100 million iPhone users may have purchased the iPhone 15 Pro or iPhone 15 Pro Max since they launched last year. This may seem like a large number. But then consider that 100 million out of 1.46 billion total iPhones in use equates to only 7% of the installed base. And then, on top of that, consider that many iPhone 15 Pro Max/Pro users do not use U.S. English on their devices, and may not be comfortable switching to it. That suggests that probably at least 95% of iPhone users won’t be able to use Apple Intelligence at launch, unless they buy a newer model and use U.S. English.
All Macs and iPads with Apple’s M-series (“Apple silicon”) processors will be able to run Apple Intelligence. Macs with M1 processors debuted in late 2020, and some iPads with M1 processors started shipping in May 2021. Apple has mostly sold M-series Macs since then, and those are the only processors that Apple sells in Macs today. So you could have a Mac that’s more than three years old that will get to use Apple Intelligence features. However, the situation with iPads is a bit different; M-series processors were only included in iPad Pro models and the iPad Air (5th generation).
For the iPhone, no models shipped prior to the fall of 2023 will be able to run these new features. Not only that, but a substantial share of iPhones sold within the past nine months (i.e. the non-Pro version of iPhone 15, as well as older models that Apple still sells) won’t be able to run these features either.
When the iPhone 14 Pro and iPhone 14 Pro Max were announced in September 2022, Apple said that it was “Powered by A16 Bionic, the fastest chip ever in a smartphone,” and that it contained “a new 16-core Neural Engine capable of nearly 17 trillion operations per second.” For years, Apple has been touting this Neural Engine, which is the main part of the chip that handles AI features, telling us how powerful it is. Yet now, less than two years after its release, the flagship iPhone 14 Pro cannot run Apple’s new marquee features. And the nine-month-old iPhone 15 (non-Pro) uses the same A16 Bionic as the iPhone 14 Pro and Pro Max.
The iPhone 15 Pro’s A17 Pro CPU is, according to Apple, “up to 10 percent faster … and the Neural Engine is now up to 2x faster.” The neural engine was previously only in iPhone and iPad chips, and was added to the Mac with the M1 processor; you’d expect the processors on mobile devices of the same age to have just as powerful capabilities.
People often forget about the Apple HomePod smart speaker; it’s far from a top-selling item. The most recent model was released in January 2023. Powered by the S7 chip (a hand-me-down from the Apple Watch Series 7), its Neural Engine simply isn’t powerful enough to run Apple Intelligence-enhanced Siri on-device.
Perhaps Apple could make an exception to allow the HomePod to have AI Siri commands processed via its Private Cloud Compute data center. However, unless Apple did the same for users of older iPhones, device owners would be further convinced that Apple was trying to force them to buy new hardware.
Apple may eventually release an updated HomePod with AI-enhanced Siri. Only time will tell.
Something in pre-iPhone 15 Pro models may be holding these devices back, preventing them from using Apple’s newest features. (Or, at least, Apple thinks everyone will make that assumption.) One difference in these devices is the amount of RAM they have; 8 GB compared to 6 GB for the iPhone 14 Pro and the iPhone 15. All M-series Macs and iPads have at least 8 GB RAM. Apple has been touting the speed of the neural engine for so many years, but could RAM be a differentiating factor, too?
Another possibility is the GPU performance of different devices. One metric for GPU performance is TFLOPS (trillions of floating point operations per second). The M1 processor performs 2.290 TFLOPS; all other M-series processors are faster. The iPhone 15 Pro only gets 2.147 TFLOPS, and the iPhone 14 Pro 1.789 TFLOPS. Could the cutoff for acceptable performance be around 2 TFLOPS plus 8 GB RAM?
One of Apple’s key selling points of these AI features is that most of the processing occurs on-device. Apple says that this helps maintain your privacy. However, it comes at a cost; on-device processing requires a lot of computing power. Given that Apple has been building on machine learning for years—mainly around Siri, but also for many other tasks—it seems that the company did not plan for these features until after the design of the iPhone 14 Pro was finalized. This seems like quite the oversight on Apple’s part, if that’s the case. Were they caught flatfooted by the sudden boom in generative AI, that hit the mainstream in late 2022 with the launch of ChatGPT?
Apple Intelligence is rolling out in the fall “in beta.” This likely means that it won’t be considered final for several months (if not for a year). To some, it may look like Apple is trying to force people who didn’t buy this year’s Pro iPhone to upgrade to get these new features, which Apple calls “AI for the rest of us.” It’s disturbing to think that a barely two-year-old flagship iPhone is already out of date—and that non-Pro models of the current iPhone are also not supported. Would Apple dare sell a base iPhone 16 a few months from now without support for Apple Intelligence? What about the iPhone SE (4th generation) that’s expected to launch early next year? Apple could choose to use AI to further differentiate between its lower-end and higher-end iPhone models, forcing power users to pay for Pro models to get more advanced features.
AI for the rest of us? Not really; at least not yet. For now, it’s more like “AI for very few of us.”
You can also subscribe to our e-mail newsletter and keep an eye here on The Mac Security Blog for the latest Apple security and privacy news. And don’t forget to follow Intego on your favorite social media channels: 
