How Artificial Intelligence is Transforming Mobile Technology

Dec 24, 2018
Vote on Hacker News

selfie on mountain

 

In today’s smartphone market, competing technologies are under substantial pressure to establish a lasting innovation advantage which will make their phone ‘smarter’ than rivals’. It’s not easy to stand out.

Whatever ardent Android and iPhone platform users say, most phones in the same price segment come equipped with much the same hardware and software which does much the same thing.

Have we reached ‘peak smartphone’? Smartphone sales have slowed as the market has matured.
Smartphones are following the long established product lifecycle.

 

The saturation level of product maturity

 

figure of Apple's plateaued sales
Sales of Apple’s flagship iPhone product range are a case in hand. It appears Apple have reached ‘peak smartphone’

 

Adding to this pressure between Android and Apple, Statistica data suggests that we’ve reached the saturation level of product maturity now. Simply put, everyone who wants a smartphone has one. The industry calls this ‘Peak Smartphone’, a trend underpinned by a number of overlapping consumer behavioral traits, primary among them individuals holding on to their phones longer. For instance, iPhone tenure has risen from 24 to 30 months in just the last 2 years. There was also a reduction in the number of people purchasing phones under contract (which more naturally led to an upgrade at the end of each contract cycle).

 

Artificial Intelligence (AI) to the rescue

How then, can phone manufacturers differentiate themselves – when they so clearly need to, in a plateaued marker?

Software and the dedicated AI hardware required to run it efficiently is one way. Investments in the field of Artificial Intelligence in mobile phones has created a new ‘battlefront’ in the fight for customers.  As such, many phone makers are queuing up to adopt Artificial Intelligence in an attempt to reinvent and deliver a more compelling user experience.

In fact, pulling back the curtains and taking a look at Apple’s iPhone Keynote back in September, AI is the ‘hidden’ feature behind nearly their entire roadmap. The 2018 iPhone XS had a second generation A12 ‘Neural engine’ processor, a camera built entirely around AI technology (I explain below for details how it works) and facial recognition facilities which run on AI. Critically, Apple has opened these new AI facilities up to their developer community, to build an AI-based ecosystem around their core iPhone product, something Apple typically does when it’s serious about a strategy.

 

Not just Apple

It’s not just Apple investing in these AI steps. AI has been the main ingredient powering several smartphones launched this year. In September 2017, HUAWEI revealed that its Kirin 970 chip has built-in AI capabilities via the neural processing unit. The company has also pitched an open platform for mobile AI which they hope will become a commonly accepted standard Samsung is also working on adding AI-specific core to their mobile phones. According to news published in the Korea Herald, Samsung phones will, from now on, have dedicated chips that will improve AI processing. Google’s CEO summarized the situation from their point of view at the Google I/O 2017 keynote suggesting that the company is undergoing a shift from mobile first world to the AI world.  Their ‘Pixel 3’ release this year includes a host of AI-based features, not least a new generation of the OK Google Assistant.

AI is the defining technology of our time, and mobile represents our favorite way to interface with the internet. Here’s how AI and mobile overlap, starting with smartphone capabilities and covering some broader applications of both technologies.

 

AI in your phone’s camera

Starting with the relatively benign end of the pack, AI is now the bulk of one of the most used parts of your phone – it’s the camera. It is no exaggeration to say that the ‘cameras’ in new phones do not take pictures. They gather all of the available image data, store it and interpret it, according to the user’s requirements, after the shutter is pressed.

The new iPhone XS, for example, allows the user to adjust the Depth Of Field after the image has been taken. The ‘Top Shot’ feature on this year’s Google Pixel 3, lets the phone capture multiple images and then uses image recognition to determine which is the best one. Photobooth, another feature in the camera stable on the same device, takes selfies and pictures of you with your friends automatically (with no button being pushed) but only when it senses that you’re pulling funny faces.

Many Android phones automatically collate collections of pictures and videos taken on a particular day or location into an animated ‘story’ with automatically overlaid music.

 

Artificial Intelligence Is Driving Mobile App Personalization

If there is a key use case for Artificial Intelligence in the consumer realm, it is the personalization of the user experience that people have in using a digital product. The specifics of the way you interface with your technology are unique. Where you look, what you click, the tunes and TV shows you like and the order in which you do things are all, just like your mum told you, perfectly you.

The rate at which AI is gaining momentum is derived, in large part, from its ability to perform this important task. When Amazon offers you products that people like you bought, Netflix tells you there’s a 97% chance you’ll like this Romantic Comedy or the BBC iPlayer puts the news at the top of your recommended viewing list – AI is behind those personalizations.

The future of mobile app personalization from AI will involve even more variables, derived from the data that is captured by the many sensors on your phones. Technologists are working on the incorporation of context-aware sensors, voice search, face recognition, chatbots, and a lot more, for the future.

 

Artificial Intelligence and Internet Of Things

The Internet Of Things is the collective name given to internet connected assets – enabled with cheap sensors and (typically) small amounts of processing power. The concept applies to everything from connected traffic lights, to buses which inform the bus stops ahead of them about the time it will take them to get there. It includes smart-watches and even children’s backpacks. It is no exaggeration to say that almost any consumer product being produced these days contains an internet connection and that soon, any asset worth $50, or more, from cows to cathedrals will be online.

All of these devices generate streams of data and in that data lie the solutions to many of the problems we face but can’t solve. There’s an analogy that’s used to explain the value of the data created by the industrial internet. Imagine you dropped your keys on a walk home from dinner at a friend’s house one night, after dark. You look behind you and see a street, lit with lighting on one side. Where would you look for your keys? It only makes sense to look in the lighted area – you won’t see them anywhere else. But the keys could just as easily lie out of sight, in a darkened corner which the light does not reach.

To this point, we have simply accepted hundreds of inefficiencies every day because we have had no alternative. The solutions to those inefficiencies were not available to us – just as the light didn’t fall on the whole street. When you waited in your car at a red light on a street with no other vehicles around, when your plane was delayed because of an engine fault no one saw coming and when your lettuce cost a little more than it should because the farmer had to use extra fertilizer to ensure the whole crop was covered – you’ve suffered as a result of those inefficiencies.

The Internet Of Things will connect each of these mobile assets (vehicles, plane engines, crop fertilizers) to the world wide web. AI will help us establish the patterns in the data they generate – assisting us in sorting the wheat from the chaff.

Analysis of Rolls Royce jet engines, for example, has established data streams which precede substantial failures, allowing engineers to conduct proactive maintenance and reduce downtime. Swarm computing (where, for example, multiple vehicles interact with traffic lights to produce a more efficient traffic flow) will reduce commute times. Association of the locations in which fertilizers were applied and the resulting yields will mean that farmers can cut back where less is required.

 

Artificial Intelligence and Mobile App Authentication

Artificial Intelligence is also having a very real impact on a variety of cybersecurity measures. Unfortunately, AI is not just used by ‘the good guys.’

Security is now one of the biggest concerns for the authors of any computer software. Artificial Intelligence is used here, too, to gather, process and interpret the data seen by computer systems prone to attach and identify likely threats. Cyber-security software examines data streams for telltale signs such as location, device identity, to enhance the reliability of the authentication which takes place on mobile apps. Some can even use social media profiles and posts

Today’s cutting-edge cyber security applications learn more about mobile user behavior and will work on other essential security measures. Tangerine Bank, for example, has already combined authentication with assistance so that users can interact with their mobile apps using iris and voice recognition.

But hackers have AI at their disposal, too. Examples include adding special overlaps to images so they are misclassified by security systems – effectively making them blind (while still appearing to work.)

Bringing it all together

It’s not hard to see AI as the glue tying together the user interfaces on all of our internet interactions at some point in the near future. It already operates for us, behind the scenes, in ways that are not obvious to most technology consumers but which are becoming more mainstream by the day.

Simple personalization has been around for years – in those Amazon recommendations. Facial recognition and entertaining personal stories concocted from image libraries are – frankly – when examined, remarkable but are now commonly accepted as a day to day occurrence. We are almost at the stage at which conversations with Siri and OK Google Virtual Assistants are as fluid as those with a real person.

Now, with the entire mobile industry betting their not insubstantial resources on open access to the entire development community to hardware processors specifically designed for the task, every app we use and every (mobile) physical item we interact with will be first connected to the internet, second analyzed and third optimized to achieve our intent. AI is transforming mobile technology, everything from smartphones to jet engines – often without us being aware.

 

 

Ralf is an I.T expert and a technology blogger, he writes about mobile phones and the latest technology news.  He currently works at Whatphone.com.au as a content manager and his writings can bee seen on various technology blogs. He also loves taking pictures when free. You can follow him at Twitter @IamRalf12.