Usually known for its secrecy, Apple Inc. has been relatively vocal about a key aspect of its future over the past few months. Through acquisitions, investments, research and public comments, Apple has been trying to send the message that if artificial intelligence is indeed the future, the most valuable company in the world is getting prepared.
has been much quieter, however, about what exactly it intends to do with the AI knowledge and experience it is gaining. Perhaps the most obvious outward example of its artificial-intelligence applications, the Siri voice assistant, inspires mixed feelings from users, especially those who have welcomed Amazon.com Inc.’s
Alexa into their homes. And while AI is central to Google’s core ad and search businesses, Apple’s main business is still making iPhones, where AI remains a small component.
That’s not to say the company doesn’t have big plans in store. The Apple of the not-too-distant future may be able to diagnose a litany of complex health problems before you even know to get tested for them. Its software could help self-driving cars avoid hitting pedestrians even in a blizzard. Augmented-reality glasses, for which it Apple recently received a patent, could identify people it’s never seen before by reviewing your contacts and Facebook profile.
For now, though, Apple must walk a fine line, integrating cutting-edge new features into its products without alienating users wary of drastic changes to Apple applications to which they have become accustomed. The company is known for waiting for technology to reach a certain threshold before adopting it in full and pushing it to the masses, and Apple will likely take a similar approach with AI.
Building a smarter iPhone
has set out to be an AI company, but Gartner analyst Brian Blau said Apple “in no way” wants to be thought of that way. He believes the company will employ the “smartness message” when needed, but mostly stick with the “customer-experience story” when announcing new products and features.
AI in 2018: Google seeks to turn early focus on AI into cash
Apple Chief Executive Tim Cook, for his part, has said that users just want their products to work, and that they aren’t so concerned with what’s behind them.
“People don’t think about well-integrated machine learning,” he told the MIT Technology Review in June. “They don’t even know it’s there.”
Financially, far-out AI ventures like autonomous driving remain an “optionality” for Apple investors at this point, according to Gene Munster, a former Apple analyst who now runs AI-focused venture-capital firm Loup Ventures.
“No one’s holding Apple stock based on what they’ll do about cars,” he said, noting investors are still focused on iPhone sales and even services revenue.
Subtle applications of AI could help drive iPhone sales, however. A better camera, a better battery and a well-functioning facial-recognition system all prompt people to choose Apple’s products over those made by rivals, and AI is a driving force behind improvements in those areas.
Artificial intelligence remains crucial to Apple’s future amid a maturing smartphone market. Overall smartphone sales are projected to grow at only a 3% annualized clip through 2021, according to market-research firm IDC. Consumers are holding on to their phones for longer, and local players are gaining steam in China.
AI features provide a way for Apple to make its current lineup of products stand out and could enable the company to branch into new areas, such as automotive technology.
“Hardware is what pays the bills and continues to be the jewel in Apple’s ecosystem, but software going forward provides a key piece of differentiation,” GBH Insights analyst Daniel Ives said.
The company doesn’t disclose its spending on AI projects, but Ives estimates that the company spent some $1 billion on AI in the past few years. Loup’s Munster believes the company allocated about 5% of its research-and-development budget to AI in the past year and will do so again next year. Apple’s overall R&D expenditures totaled $11.6 billion for the fiscal year that ended in 2017.
Recent financial moves highlight Apple’s commitment to AI. The company bought music-recognition app Shazam earlier this month, reportedly for about $400 million. In a statement, Apple said that the service was “a natural fit” for Apple Music, though Apple’s long-term plans for the acquisition likely go beyond music, given Shazam’s expertise in analyzing audio.
Using natural-language-processing technology, the company may one day be able to perform instantaneous translations, for example, the way Google’s Pixel Buds have set out to do. Apple highlighted natural-language processing when it launched its Core ML developer tools earlier this year.
Even a $390 million award from Apple to supplier Finisar Corp.
has AI implications. The company makes vertical-cavity surface-emitting lasers, or VCSELs, to power the 3D-sensing technology behind Face ID, and Apple wants to ensure that it has enough of the lasers to meet iPhone X demand.
Right now, 3D-sensing is limited to the front-facing camera and makes use of neural networks to create a mathematical model of your face. But down the road, the technology could make its way to rear-facing phone cameras and eventually other Apple products. Think a HomePod speaker that senses when you come into a room and automatically dims your lights at the time you prefer.
In general, Apple is making progress with AI components of the iPhone, including its new A11 processor, which has an integrated neural engine.
“The fact that this technology is embedded in the device and not in the cloud enables a better user experience and better protects user privacy,” said Forrester analyst Thomas Husson, who deems it the company’s biggest AI advancement of the year.
The A11 Bionic chip in Apple’s iPhone 8 Plus and iPhone X enables portrait lighting, a camera feature that makes use of machine learning to identify facial features and change the lighting on your face. It’s a subtle but powerful improvement to the camera, and one that could prompt more people to choose the iPhone or upgrade from an older model.
Cars and more
Beyond the iPhone, other areas show promise. Apple seems to have given up on its hopes of designing a physical Apple Car and is now aiming to be the brains behind self-driving vehicles, something Cook recently called “the mother of all AI projects.” Autonomous-driving research is an expensive proposition, but one with the potential for a big payoff down the road. The Boston Consulting Group expects the market for partially and fully autonomous vehicles to swell to almost $77 billion by 2035, and Apple wants a piece of that pie.
AI in 2018: Waiting for driverless cars to become a reality
A November research paper authored by two Apple employees detailed breakthroughs that the company has made in using Lidar sensors to detect pedestrians, cyclists and cars. Apple’s director of AI research, Ruslan Salakhutdinov, went into more detail about the company’s efforts at a December event for the broader AI community. He highlighted progress the company had made in using sensing to detect pedestrians who were partially obscured by parked cars and training camera sensors to recognize raindrops on a lens. The company is also working on 3D maps that could help autonomous vehicles navigate their surroundings.
“It’s clear that AI is helping under the covers in a lot of areas we didn’t know about before,” Gartner’s Blau said of Apple’s latest disclosures.
Apple is still in the early stages with its self-driving ambitions, as are its rivals, but the company is farther along in its health-technology work.
“Health is fertile ground for AI,” IDC analyst Ramon Llamas said. “Once you collect information about millions of people, you need something to make sense of it.”
Apple recently announced a study that uses deep neural networks to analyze heart-rate data from the Apple Watch and detect atrial fibrillation, a form of irregular heartbeat. Future versions of the watch could introduce new biometric sensors, including those that measure blood-glucose levels without breaking the skin or detect your temperature, which would open up more diagnostic opportunities.
Wearables that can not only track what you’ve done but also analyze patterns and suggest behaviors hold particular appeal in the health industry. Smarter AI applications within the Apple Watch could help boost sales of the device among insurers and corporate wellness programs. The more interest in the Apple Watch that the company can generate from the health and medical communities, the less reliant it will have to be on device sales to ordinary consumers, something rival Fitbit Inc.
is also realizing.
Just as the Apple Watch is starting to tell people information they didn’t know to ask for, Siri must too get smarter and move from merely telling folks about appointments on their calendars. A smarter Siri will be able to think for you, to analyze your behaviors and remind you to do things you never thought you needed. It could one day tell you to decline an invitation to lunch at a particular restaurant because your Health app indicates you got heartburn there last time.
AI is an “arms race,” said GBH’s Ives, and Apple must work to catch up with Alphabet. Google CEO Sundar Pichai has spoken regularly of the company’s transition from mobile-first to AI-first, and its Waymo self-driving technology is widely considered the most advanced.
Apple may be playing from behind, but in once sense it has time on its side. Loup’s Munster points out that Apple was an “AI company” back in 1993, when it employed handwriting-recognition technology on its Newton digital assistant before Google and Amazon
had even been founded.
The Newton, however, was one of the gadgets that taught Apple to always perfect technology before pushing it out to the general public. In the new wave of AI, Apple will likely take that lesson to heart and make sure the interesting developments happening in its labs only make it out of Apple Park when they’re ready to be big sellers.