Apple could bring generative AI to iOS 18, but is it too late?

If you’ve ever used Siri on an iPhone, iPad, Mac, or even an Apple Watch, you’re probably already very familiar with its failings. But there are signs that Apple is getting ready to improve matters by venturing into the new and exciting world of Large Language Models and generative AI. The only question is whether it’s too late to a party that already has ChatGPT and others.

If Apple really does intend to make use of generative AI it will surely go beyond just improving Siri. Rumours have already said as much, but it’s Siri that will touch so many people’s lives on a daily basis. In a world where Alexa is powered by AI and companies like Google offer hugely impressive AI-powered features, Apple’s starting late. And with so much catching up to do, does it have the know-how required to do it?

Apple’s plans for the future

Let’s back up a minute. I mentioned there were signs that Apple is starting to get interested in generative AI and those signs come via Haitong International Securities analyst Jeff Pu.

Writing in a research note seen by MacRumors, Pu says that his supply chain checks point to the fact that Apple is set to build hundreds of AI servers before 2023 is out. The note continues, saying that Apple will significantly increase that number next year. However, not all of Apple’s future generative AI chops will rely on the cloud – the MacRumors report notes that Apple will offer a combination of cloud-based services and so-called “edge AI.” That’s a term given to services that run on local devices more than in the cloud – in Apple’s case, that will mean AI that runs on iPhones, iPads, Macs, and even Apple Watches. That approach would likely allow Apple to continue to put user privacy first by keeping data on-device rather than offloading it to servers in data centres.

In terms of timeframes, it appears that Apple will work to get its new initiative up and running towards the end of 2024. If so, it’s possible that Apple is aiming for a launch alongside the iPhone 16. Older devices may also benefit via the iOS 18 update, likely to be announced in June 2024 before shipping to the public the following September. In fact, Pu mentioned that the AI update would likely arrive as part of a software update, making iOS 18 the obvious bet.

If that is indeed the case, it can’t come soon enough. Because Apple needs it. That isn’t to say that Apple can’t do AI, because it can. It’s pretty good at it, too. The A- and M-series chips that power iPhones, iPads, and Macs all feature a Neural Engine component designed specifically to help with AI processing workflows. And those workflows are put to good use, too. Modern iPhones take some impressive photos all thanks to the AI used in their photography and videography pipelines. But that’s a different kind of AI. Apple’s great at it. But generative AI is a different beast. And right now, Apple’s a long way behind the competition.

So much catching up to do

Even the most ardent iPhone-toting Apple fan has to know that the competition is in some ways lightyears ahead. Siri is the obvious example, with Alexa and Google Assistant able to understand what is asked of them more easily before going on to handle that request with ever-increasing aplomb. By contrast, Siri still blights so-called smart speakers like the HomePod and HomePod mini. Those speakers are excellent buys in many respects, but that’s in spite of Siri, not because of it. If they didn’t sound so good, they’d be impossible to recommend.

Pocket-lint

But it isn’t just the world of digital assistants that Apple finds itself behind the eight-ball. A glance at the recent Pixel 8 and Pixel 8 Pro releases shows that Google is leveraging its AI smarts to create features the likes of which Apple can’t quite match. Features like the Magic Editor, a tool that will have you questioning what a photo even is anymore. The new Pixels can take components of images and remove them, add new ones, change existing ones, and just generally fiddle with images in ways you wouldn’t even imagine. It does it all using the power of generative AI, adding content to images based on what’s already there. And it is, frankly, astonishing what it can do.

Where’s that on the iPhone? Nowhere, although it’s surely only a matter of time before the App Store is full of apps willing to do this kind of thing while Apple gets its act together.

The Siri situation

It’s impossible to talk about AI without the most obvious part of Apple’s ecosystem that would stand to benefit from a helping hand.

Siri has been around since 2011 after it debuted with the arrival of the iPhone 4S. And while it would be unkind to say that it hasn’t changed much since then – it absolutely has – it hasn’t come on in the leaps and bounds that you might expect from a dozen years of development.

In fact, at times, Siri can be downright terrible. Its inability to answer the most simple of questions is regularly mocked on social media. You could forgive Siri if it struggled with complex questions, but sometimes it’s all about the phrasing of the question – use a phrase it isn’t expecting, and the answer can be terribly wide of the mark.

Any improvement there would be most welcome, and if it takes a few hundred new servers to make it happen, so be it.

Thankfully, it would appear that Apple has its sights set higher than just making Siri less difficult. A recent report by The Information hinted that Apple’s AI plans go way behind Siri. Talk of giving people the power to automate complex tasks will warm the heart of many a power user, especially if that spills over into the often hamstrung Shortcuts app.

But all of this is guesswork at this point, and Apple might just be building new server farms to let us use AI-generated wallpapers on our Lock Screens. Let us hope that’s not the case, though.