Apple CEO Tim Cook has finally elaborated on his company’s plans in the self-driving car area in an interview with Bloomberg TV, calling it “the mother of all AI problems.” And at WWDC last week the terms “machine learning,” “neural networks,” and “computer vision” were sprinkled liberally throughout the many software and hardware announcements.
It’s not hard to miss that Apple is talking a lot more about AI in 2017 than it did in 2016. It’s not because the company was throwing fewer resources at its AI work last year.
What changed between 2016 and 2017 is Apple’s marketing and PR strategy around AI and related technologies. In fact, it made a conscious decision to start talking more about its AI efforts last summer, just before the publication of Steven Levy’s behind-the-scenes look at Apple AI in August. The company has been more vocal about its AI efforts–and the people behind them–since then.
The way Apple presents its products to the world is arguably as important to the company’s bottom line as the features in the products (which more often enter established markets than create new ones). The messaging helps people feel okay about paying premium prices for Apple stuff. The marketing and PR people may be the ones most responsible for establishing and defending Apple’s famously high profit margins.
Now AI is becoming an important part of the messaging. Until pretty recently, Apple liked to avoid saying too much about the nerdy stuff, preferring to let cool new AI-powered features speak for themselves with their smarts and utility.
But, as Creative Strategies analyst Carolina Milanesi pointed out to me after the WWDC keynote, consumers weren’t connecting the dots between the features and the AI. And with Google, Facebook, Microsoft and others talking loudly about the technology inside their products, it’s become important that Apple not look like a laggard.
The AI team at Apple operates in much the same way as other service-oriented groups within Apple devoted to matters such as industrial design and security. It works with other teams across the company, adding brain power to hardware, software, and services.
You can get some idea of the breadth of AI’s influence by listening closely to the contexts in which Apple people referenced machine learning and other AI tools at WWDC. During the 120-minute-plus keynote presentation, they described how machine learning is used to inform the personalized data displayed in the new Siri Apple Watch Face, how it’s used to help Safari detect and prevent unwanted web tracking, and how devices use it to extend battery life by predicting usage patterns.
Machine learning, Apple said, is used in the iPad to detect (and ignore) when an Apple Pencil user’s palm hits the touch screen, and to recognize a Pencil user’s handwriting and index it in Spotlight. Computer vision helps the Photos app to detect places, objects, and user interests.
Apple also talked a lot about enabling machine learning work by third-party developers. It announced two new AI APIs that will let developers build Apple’s own computer vision and natural language chops into their own products. Computer vision will be brought to bare in the augmented reality experiences developers create with Apple’s new ARKit tools. The company also talked about how the powerful graphics processors in new Macs support the computationally intense calculations necessary for machine learning.
This Isn’t A Horse Race
The conclusion of WWDC 2017 last Friday marked the end of the season of developer conferences. Over the past few months, we heard Facebook, Microsoft, Google, and Apple talk at length about their own flavors of AI. After witnessing all this, some people came away thinking Apple is still trailing its rivals in AI and its applications.
I’m growing increasingly skeptical of this point of view. Not because I believe Apple AI is necessarily getting a bad rap, but because I don’t believe you can look at AI as a horse race. All the above-mentioned companies (and Amazon and IBM and others) have very different takes on AI and different uses for it, and they’re all coming at it from slightly different strategic directions. It’s Apple and a bunch of oranges.
Also, “machine learning” and “computer vision” have become marketing buzzwords. Their use in corporate messaging may distort or overplay a company’s real work with the technology, misrepresenting a company’s real investment in AI. Should IBM refer to itself as an “artificial intelligence company” or is that just an aspirational wrapper for its refocusing efforts?
Apple’s original take on AI–that we should focus on useful features rather than obsessing over the technologies behind them–may be exactly the one consumers (and journalists) should adopt.