AI is coming to the iPhone–and it could change everything

Macworld

After years of the market complaining that Apple is “behind” on artificial intelligence, the company is poised to make a big push in the technology with its platform updates this year. In a rare move, that’s been confirmed by no less than CEO Tim Cook himself, who said in the company’s most recent financial results call that the company would “share the details of our ongoing work in that space later this year.”

Of course, the company’s not really a stranger to this space: Apple has spent plenty of time deploying machine learning technology in a variety of areas for years, from photography to autocorrect. But the industry’s focus of late is on generative AI, the technology that underlies the products that have captured the zeitgeist, from chatbots like ChatGPT to image creation tools like Dall-E and Stable Diffusion.

The big questions that hover over all of this are how Apple will bring those technologies into its existing operating systems and what choices it will make in rolling them out. The company tends to be on the judicious side when it comes to deploying new features, but there are still plenty of places on its platforms where generative AI—contentious as it may be—might find a foothold with users.

Get this conversation started

If there’s an obvious place for generative AI it’s what is often considered the heart of intelligence on Apple’s platforms: Siri. The virtual assistant turns 13 later this year, and in many ways, it has not advanced appreciably from its earliest days. I’ve advocated for a Siri 2.0 for almost a decade, if not more, and generative AI feels like the opportunity to finally put that into fruition.

And it doesn’t even necessarily have to be full of all the whiz-bang capabilities that ChatGPT or Copilot have to offer; I’d settle for a virtual assistant that reliably understands what I’m asking for, or can handle two instructions at once. One of the strengths of tools like ChatGPT is its conversational manner, which is good enough that it can make you feel like you’re interacting with a person rather than a machine. Siri might have the trappings of that with its voice models, but you get enough canned responses to let the bits shine through.

Siri is also the locus of a lot of Apple’s other intelligent features, including search and interaction with other services both on the device (texts, calendars, email, etc.) and off. Apple, with its dedication to bringing processing to the device itself, may be able to leverage the interaction with users’ data stores, but it does still raise the question of what happens when you need to deal with more than just the data on a device.

Cloudy with a chance of AI

That on-device processing isn’t anything to sneeze at: Apple has included its Neural Engine, a specialized piece of hardware designed explicitly for machine learning tasks, as part of its chips since 2017.

Still, as tuned as those might be for machine learning tasks, there are some things that might simply necessitate more horsepower than a chip in an iPhone or even a professional desktop can handle.




A recent report said that Apple is discussing the possibility of licensing Google’s Gemini AI technology.



A recent report said that Apple is discussing the possibility of licensing Google’s Gemini AI technology.Google

A recent report said that Apple is discussing the possibility of licensing Google’s Gemini AI technology.Google


Google

That’s one reason why the news this week that Apple had entered into discussions with Google and OpenAI for the use of their technology makes sense. For better or worse, those cloud-based technologies offer a lot more power, especially when it comes to generative features—and, given that, it also means offloading potentially battery-consuming features.

But it’s likely that such technology is just part of a suite of solutions. It’s not as if Apple has been standing still on the AI front: just this week, it published a research paper on training AI models. While Apple’s AI tech may be best suited for specific cases dealing with its own devices, the company has been working on generative features for less time than competitors, and bringing in those sources to complement the company’s own in-house developments provides the best of both worlds for users.

AI everywhere

Perhaps the biggest thing I’m hoping to see from Apple’s AI ambitions is just that: ambition. There are tons of places throughout the company’s platforms that could benefit from AI. Whether it’s summarizing an email message (and perhaps using that summary to help inform spam filtering), improving image editing (by, say, letting you remove unwanted people or objects from a picture), or helping developers (from those who write apps to those like myself who just dabble) generate code.

I realize that AI is the current buzzword of the industry, and everything and its cousin is bragging that it’s powered by artificial intelligence these days, but this is where I’m hoping Apple can bring some sanity to the process. There are very legitimate uses for the technology and Apple has the opportunity to come up with ideas that none of us can foresee. That’s one of the places that the company traditionally excels, providing capabilities that seem at once obvious but also provide a potentially huge impact on the way we all interact with our devices.

Apple Inc, iOS, MacOS