Skip to main content

Apple wants you to know it already does great AI — but it’s ‘subtle'

Apple wants you to know it already does great AI — but it’s ‘subtle'

Share this story

Vlad Savov

Apple wants you to know it’s been working on AI for years now — you just didn’t know it. In a new feature by Stephen Levy in Backchannel, some of the company’s top execs and machine learning experts hammer home this message, pointing out all the ways in which AI is used in Apple’s products today. But, they also point out that artificial intelligence isn’t the "final frontier" for tech products "despite what other companies say."

Apple has cause to push this message. The perception of the iPhone-maker in the wider AI community is that it’s been behind the game, despite having launched Siri, its virtual assistant, back in 2011. From Amazon’s Alexa to Facebook’s chatbots, from Microsoft’s conversational computing to Google’s Allo app, the big story of tech in the last year has been the mainstream accessibility of artificial intelligence. How can machine learning make products faster, smarter, and easier to use? Everyone seems to be at it, but Apple isn’t often involved in the narrative.

This isn’t completely unfounded. Apple doesn't have a dedicated AI division, it doesn't publish research, and its emphasis on user privacy means it has less data to train its machine learning systems (although it’s working around this). Siri has gotten much better, yes, but in its early years it was often seen — albeit in affectionate fashion by loyal Apple users — as a fairly erratic virtual assistant. This perception persists.

But that doesn’t mean Apple isn’t working on it, and this is the message the company pushes in the Backchannel piece. Levy notes that back in 2014, Apple moved Siri's voice recognition onto a system based on neural networks (the building blocks of many AI techniques), but that the iPhone-maker didn't publicize it as it didn't want to signal its intentions to competitors. Since then, writes Levy, Apple has been slowly layering AI and deep learning into more and more of its products.

Here's Levy writing on Apple’s hidden, ubiquitous, AI work:

If you’re an iPhone user, you’ve come across Apple’s AI, and not just in Siri’s improved acumen in figuring out what you ask of her. You see it when the phone identifies a caller who isn’t in your contact list (but did email you recently). Or when you swipe on your screen to get a shortlist of the apps that you are most likely to open next. Or when you get a reminder of an appointment that you never got around to putting into your calendar. Or when a map location pops up for the hotel you’ve reserved, before you type it in. Or when the phone points you to where you parked your car, even though you never asked it to. These are all techniques either made possible or greatly enhanced by Apple’s adoption of deep learning and neural nets.

In short, Apple's argument is that you’ve been using AI on your iPhone — you just might not know it. These serendipitous little moments aren’t just serendipity, but the result of machine learning and "neural nets." (Although, in my experience, these moments don’t tend to happen that often. Perhaps that's just because my lifestyle doesn't include enough international flights and restaurant bookings.) Apple also says it’s not just services that are benefiting from AI, and Levy notes that even the palm rejection feature on the Apple Pencil was trained using a machine learning model. That’s right, it’s an AI pencil.

But at the same time as it says it’s really great at AI, Apple also makes the point that it’s not the be all and end all. The view in Cupertino, writes Levy, is that machine learning is only "the latest in a steady flow of groundbreaking technologies." As Apple exec Eddy Cue says: "It’s not like there weren’t other technologies over the years that have been instrumental in changing the way we interact with devices."

The company tells Levy that it doesn't have a single, centralized department for machine learning; it won't say how many people at Apple are working on ML (beyond "a lot"); and although the company's argument that its scientists don't publish their work because they're more focused on making useable products makes sense, it would hold more weight if there were more obvious products.

It’s not entirely surprising that Apple would take this somewhat contradictory we’re-the-best-but-we-don’t-care approach here. Its rivals are getting a lot of attention in this area, but Apple is known for its strict adherence to secrecy (and also, user privacy). Of course, if the company really wants to convince the world that it’s doing the best machine learning and AI work out there, the way to do it is to show it.