Over the past few years, the biggest players in tech—Apple, Google, Microsoft, and Facebook, among others—have been pivoting towards predictive smartphone technologies that give you the information you want when you want it. Their approaches have sometimes hit but more often missed. But at the Google I/O conference earlier this year, Google made it clear that it sees the future of mobile in Artificial Intelligence. And at the WWDC conference a few weeks after, Apple showed that it thought the same. For better or worse, AI is coming to your pocket.
As far as operating systems go, things are plateauing. The situation now is unlike that of a few years ago, when iOS was the go-to platform for user experience and Android was the go-to platform for tinkering. If you look at a new iPhone or Android, the functions are more or less the same, performance is smooth, and customization is solid; differences still exist, of course, but the experiences are converging.
So competition has focused on AI which gives companies new chances to innovate and distinguish themselves. Artificial Intelligence is a broad term, and in a mobile context, it generally refers to predictive software that can passively learn and independently act based on your usage or your information. The most famous examples are probably Apple’s Siri and Google’s Google Now, but Amazon, Facebook, and Microsoft have jumped on-board with cross-platform offerings (Alexa, M, and Cortana, respectively).
Predicting and Processing
Apple’s Siri (iOS 10) and Google Now/Assistant (Android Nougat) are both AI centered around your lifestyle, meaning they’ll use machine and contextual learning in order to figure out your habits. So in order to be actually useful, as the companies intend them, these mobile AI need access to a lot of your information.
Take Google Now/Assistant (Assistant is the conversational upgrade for Now in Android Nougat), which has a slew of great features. Among other things, Google Now can give you schedule based reminders to pick up groceries, scan your emails for appointments or boarding passes, and give you recommendations for music and articles based on what music you’ve listened to and what sites you’ve visited. But in order to access all these things, to have this virtual valet, you would need to give up your location, emails, and search history to Google’s cloud for processing.
Apple’s Siri has been aiming for more predictive capabilities as well, slowly shaking off its reputation as a fun search toy. Last year’s iOS 9 added a proactive Spotlight page, which tried giving news and location-based suggestions in the style of Google Now. This year’s iOS 10 upgrades include features like smarter messaging suggestions on your keyboard and facial recognition. Though it’s still accessing your messages and photos, Apple has promised to keep all the processing on device, as opposed to sending it off like Google Now.
Two big concerns often associated with artificial intelligence are privacy and existential danger. The latter doesn’t apply much for the low-level AI in smartphones (at least not yet), but the former, if you decide to use a mobile AI, will immediately involve you.
Apple’s approach is clearly more “private” for the user than Google’s, but we should keep in mind that on-device processing may have several big drawbacks that we won’t entirely know until Apple adds more complex AI functions to Apple, which it will in order to compete with features Google Now already has. The first concern is speed--how quickly can AI processing be done on the phone versus a speedy over-the-cloud mainframe? The second concern is battery life—if the processor is calculating quickly enough, does it use too much battery? The third, and most important concern, is functionality—how useful can Siri actually be as an AI if it limits so much processing to the phone itself?
Currently, Google Now stands as the more functional mobile AI. It’s cloud-based approach addresses all three of the concerns I listed above simply because it offsets its processing to “somewhere over the rainbow”. But of course, as Now grows more functional, it follows that it will start needing more information, more than the copious amount it already accesses. And all your information, too, will be going to the same “somewhere over the rainbow”. The problem here is oft-stated but nonetheless significant: you don’t know what Google is doing, or what it can do, with the increasing amounts of data it’s getting from you. You can only trust, and sometimes, that’s not enough.
The problem is, you can’t have the functionality, can’t have the personalization, can’t have the experience of a smart assistant if you don’t give your assistant something to work with. It’s why a balance between Siri and Google Now would be so great, but so difficult to pull off as something anyone would use. And as for policy, it’s difficult for the government to regulate how a company like Apple or Google will use the information you willingly relinquish to them.
And even if it’s at least certain that Apple, say, will use your information only to fuel Siri, the information you’re getting back from Siri isn’t really in your hands; it’s still in Apple’s. By omission and recommendation, Siri could theoretically alter your life pattern to prefer certain restaurants or certain news topics while ignoring other restaurants or topics, simply because you aren’t aware of them. Suggestion is powerful, so it’s easy to see where this could go wrong.
To that end, I can only make one recommendation: educate yourself. Selectively research and choose the AI services you want and only give up the information you feel comfortable giving up. Balance what you give away and what great things that Apple or Google or whatever company will give you in exchange. Take advantage of the AI in your pocket, but don’t let it take too much advantage of you.
Luke is currently a freshman at Harvard College interested in the social sciences. He avidly follows news about smartphones and developments in mobile tech. In high school, he wrote a research essay on the constitutionality of NSA Internet surveillance; after this experience, he began becoming more aware of the complications between Internet technologies and policy. These interests led him to the Future Society.