The AI upgrade cycle is here

Photo by Vjeran Pavic / The Verge

The new Apple Intelligence features coming to iOS 18 could be the most impressive integrations of AI software and consumer hardware to date. Apple’s AI tools will be able to take actions based on what they know about you, manage your notifications, and rewrite text. They’re the kinds of things that could make day-to-day use of your iPhone a lot better. But they won’t be available unless you have one of Apple’s latest and most expensive iPhone models.

AI has quickly become the latest entry in the tech industry’s never ending desire to drive an upgrade cycle. A few years ago, every smartphone maker raced to 5G; more than a decade ago, the TV industry pushed for 3D TVs. Right now, every tech company clearly sees an opportunity with AI and is adding AI features confined to their latest and greatest devices as a result. But like the race to 5G, the mad rush toward AI is happening quickly and before the tech has been proven useful and its problems ironed out.

Image: Apple

Apple Intelligence can make a “memory movie” using your photos based on a text prompt.

To use Apple Intelligence on an iPhone, you’ll need to have an iPhone 15 Pro (which starts at $999) or iPhone 15 Pro Max (which starts at $1,199). That probably comes as a disappointment to iPhone 15 and 15 Plus users, who have phones that aren’t even a year old. iPad and Mac users are luckier: you’ll need to be running one of those devices with an M1 chip or newer to try Apple Intelligence for yourself, so any device from the past few years will work. (The M-series chips, of course, have been another Apple-encouraged reason to upgrade your gear.)

Apple isn’t necessarily playing tricks here, as there is a real need for improved hardware to run advanced AI systems. AI processing is demanding. And Apple isn’t the only tech company gating some of its best new AI software features to specific hardware. Only the new Copilot Plus PCs from Microsoft and others will support some AI features in Windows like the (controversial) Recall. Google initially said its Gemini Nano model would only be available on its Pixel 8 Pro, even though the regular Pixel 8 has the same Tensor G3 chipset.

On the other hand, it’s not clear how firm those hardware limitations really are. Apple hasn’t said exactly why older iPhones can’t use Apple Intelligence. Many AI features are local, but more advanced queries are often offloaded to the cloud. And given that Google recently said that the Pixel 8 actually will be able to run Gemini Nano, the limitations may not be as cut and dry as the tech companies make it seem. Apple didn’t respond to a request for comment about the hardware restrictions.

As an iPhone 12 Mini super fan, this potential upgrade push is a huge bummer. Since its 2020 launch, the small phone has remained pretty much perfect for me even with its lackluster battery. I haven’t had any significant reason to upgrade my phone for years, saving me bundles of money while still getting to use the vast majority of new features included with iOS upgrades.

I’ve been hoping to hold onto my Mini phone until it goes kaput. But if these Apple Intelligence features actually work, I might end up getting a new iPhone much sooner than I expected.

To be fair, you don’t have to upgrade your devices just because of some new AI features. iOS 18 will be available to every iPhone that runs iOS 17 — devices going back to 2018’s iPhone XR and XS — meaning that a lot of people will still get features like improved Home Screen customization, handy new Messages tools (tapbacks with any emoji!), and the Passwords app. And you can of course keep on using your current Windows machine and Pixel phones if they’re serving you well.

But you can be sure that Apple, Microsoft, Google, and others are going to continue to pack their operating systems with new AI features that require better hardware. Clearly, they think (or maybe just hope) that the features will be compelling enough to make an upgrade a lot more enticing, even if you don’t really need it.

Posted from: this blog via Microsoft Power Automate.