The reason has to do with two toolsets Apple debuted that give developers access to superb augmented reality (AR) and AI technologies. Together, they will be game-changers for marketers looking for new content and creative experiences. That’s all of you, right?
How can Apple change the AI game when it’s so “late” to the party? After breaking from the gate first with Siri, Apple watched others catch up and pass it by.
Now, in addition to several superior Siri clones, Amazon, Facebook, Google, IBM and Microsoft have all beat Apple to market with advanced AI services. They’ve already been offering AI technologies via application programming interfaces (APIs) to enable the world’s developers to access machine learning in their clouds for anywhere from many months to a couple of years. Watson-powered chatbots are a dime a dozen, and API.ai chatbots are aren’t too far behind.
-- How many of you are actually using those other AI technologies already? I thought so. Apple’s AI may be a few months behind other tech behemoths, but it’s still early, early days.
-- The little matter of about a billion devices already enabled to use Apple’s AI and AR capabilities, about 680 million of which are iPhones.
Powerful forces have been set in motion. Developers who do that installed base math won’t think long before diving in.
Here’s the best part for marketers, the other tech behemoths, and customers: Apple won’t be the only beneficiary of the gold rush toward AI and AR being unleashed. That’s because the AI/AR world isn’t as simple as the mobile phone business or the music business. This won’t be a case of the iPod versus Diamond Rio or the iPhone versus BlackBerry.
Juxtaposed against Apple’s immense (and coherent) installed base and extreme approach to individual privacy,* the other guys have superb AI technologies of their own and proven cloud infrastructures. Also, much of their technology includes open-source software around which even the competitors collaborate. So, probably for the entire professional careers of anyone reading this column, it’ll be a multiplatform AI/AR world.
A multiplatform AI/AR world means even faster technological advancement, as the platform competitors spur each other forward — and lower prices, as they compete hard for your business.
Specifically, Apple introduced Core ML, the toolkit for developers to use to build machine learning AI capabilities into apps, and ARKit, the toolkit for developers to build augmented reality into apps.
Apple says Core ML supports deep learning operating in a way that allows it to fully exploit the abilities of the latest and fastest chips, which means maximum performance. It enables computer vision, like face detection and tracking, text detection, object tracking, etc., and natural language processing. (Separately, Apple said a Siri upgrade eliminates all preprogrammed responses. Siri’s AI will now generate everything she says on the fly.)
Apple says ARKit “fuses camera sensor data with CoreMotion data.” CoreMotion is Apple technology that provides developers with
hardware-generated information from the iPhone’s and iPad’s accelerometers, gyroscopes, pedometers, magnetometers, and barometers.
“These two inputs allow the device to sense how it moves within a room with a high degree of accuracy, and without any additional calibration,” Apple said.
What that means, according to reports from people who saw the demos, is that apps using ARKit are able to augment the real world and track it, in real time, as the user’s device moves through it. Technically, that’s hard — and a key capability needed for marketers who want to build AR-based customer experiences.
Anybody want to brainstorm cool AI-mediated AR customer experiences we might invent with these technologies? Comment below; I’m happy to host a session at my favorite Manhattan whiskey bar!
*A note about coherence and privacy: To explain what I mean by a “coherent” installed base, I’ll paraphrase Apple CEO Tim Cook’s comment last weekend that 86% of Apple’s mobile users are on its latest iOS 10, compared with 7% of Android users using the most-current version.
With regard to privacy, it’s not that Google and the others are slouches; they’re not. It’s just that Apple is kind of extreme about it. Its toolkits enable most AI processing to be done right on the device, so personal data is not transmitted. When such data is sent to the mother cloud for processing, it’s encrypted (just like the others) and scrambled together with unrelated data. Let the bad guys try to figure that out!