Artificial IntelligenceMachine Learning

Forget Siri: Machine Learning and AI is Coming For Your Smartphone

Your phone is already smart, so what’s next? Artificial intelligence and machine learning is coming to the smartphone in ways much bigger than Siri. Earlier this year, Google announced they’ll be working with semiconductor company Movidius By licensing some very special Visual Processing Unit chips, Google plans to move away from data centers and into users’ real lives—and phones.

Current smartphones already use a certain amount of AI. Siri, Cortana, Alexa, Viv, and other artificial assistants are data-based machine-learning wonders. As are several apps, including Google’s own photo-sorting app, which mines your photos’ data and sorts them in appropriate categories. Machine learning and AI have been kept relatively far away from smartphones not only because the technology is still developing, but because it devours energy. The amount of data required to perform even the smallest task is huge and quickly sending and receiving information back from data centers is simply impractical. That’s why Google’s pact with Movidius is such a big deal.

The Movidius chips will bring low-power, machine-learning processors directly to your smartphone, meaning data can be examined right there, instead of going back and forth and around the world. More importantly, this new technology is hardly a “niche.” MIT researchers recently created the similar Eyeriss, and Apple recently purchased Perceptio, a startup developing AI possibilities for phones. The integration of a chip into the system of a phone is also a big step up from a single app that performs one function. Rather than just a new bit of software, the smartphone is seeing upgraded and specialized hardware coming out for the express purpose of growing AI. By creating better devices and machine learning atmospheres, the rate at which new opportunities will appear is exponential.

Common uses for the new technology seem obvious. AI is perfect for face and object recognition as well as object tracking. Apps are already being made to process language, spoken and written, and help users take better photos, but those are the obvious uses. Machine learning will help smartphones understand their users in unprecedented ways and eventually create relationships that are much more complex than “input question, output answer.”

For example, phones may be able to understand a users’ current emotions and moods. By tracking a user’s usual patterns, a program can begin to understand exactly how much they tend to use apps, how hard they press the screen or toss the phone around. Sensors that users tend to forget even exist can actively monitor a person’s state. For advertisers, this is the ultimate opportunity. If consumers’ phones know who is tired or who is hungry, they can create even more tailored offers and promotions. For users, this could also mean a world of new apps.

A deep learning system could pull data from interactions with smartphones, and from social media data and algorithms, to create new opportunities in every part of life, specifically in the area of health tracking and life gamification. For those who use HabitRPG, Super Better, or the other ultra popular goal setting and life tracking programs, AI will boost this to a new level—one where the user can’t cheat or make mistakes. Machine learning algorithms can also be used to generate better health and life predictions for users. In a time where no one leaves the house without their phone, smartphones seem to be the ideal place for AI to take shape.

New Phones Means New Branding

As AI gets consumerized, buyers are going to forced to adapt to a new method of understanding phones. The market is largely divided by Apple lovers and the androids, like Samsung, HTC and LG. Buyers often base their decisions on which phone sports the newest and slickest UI, leaving the phones without any huge underlying differences—especially among android phones. Once phone companies begin to integrate machine learning and building AI into the phone, itself, brands are going to begin to look very different. The difference between top-notch and bottom-shelf AI is huge. It’s no longer just about speed, but capability. Namely, those top-notch systems will be a world away from the cheap knock-offs. The pervasive autocorrect function is a great example, where some programs really shine and make users lives better, and some completely fail. More importantly, different users might have differing opinions on the same program.

Even among those top brands, the actual feel and usage of their AI will differ. Many companies are already looking to remove the idea of “searching” or manually entering data, entirely. Instead, they expect AI assistants to 100% replace the idea of a computer “device” in the next decade. This means, phones will no longer be a machine, but a real first iteration of daily interactive AI.

New surveys, including a popular one from Ericsson, reinforce the fact that more and more users are becoming comfortable with the integration of AI into their daily lives. More importantly, they’re expecting it. Over 100,000 consumers spanning 40 countries were surveyed, and one in two smartphone users admitted to preferring an AI interface over devices with screens. The generation of streaming natives, especially, is looking forward to the end of the screen age. They expect homes to be embedded with sensors in the upcoming years, and those sensors better be hooked up to future smartphones.

It’s difficult to say whether the expectations of those surveyed can really be met in the next decade, but the desire to involve AI in consumers’ lives is clear. The introduction of new hardware means smartphones can begin processing more data and expand the possibilities of machine learning. Companies will have to start picking teams and choose which startups and styles to bet on. In the end, advertisers may be happy to get more opportunities to understand consumers, but it’s the users who will really benefit from the new technology. AI will be making lives easier, and plucky new startups will be showing us all the apps, tech and tools we never knew we wanted.

Like this article? Subscribe to our weekly newsletter to never miss out!

Previous post

"The application of informatics will continue to grow with access to more data and technology, enhancing human health innovation..." - Interview with Vium's Joe Betts-Lacroix

Next post

"Spark has the potential to be as transformational in the computing landscape as the emergence of Linux..." - Interview with Levyx's Reza Sadri