iPhone 7, Machine Learning And Beyond

Apple are embracing AI and machine learning to improve the iPhone


Will machine learning at Apple be a major part of the iPhone 7 launch? As this is being written, journalists and Apple fans are eagerly awaiting the unveiling of the iPhone 7. Apple face several challenges with the release of this new model, none more so than in simply maintaining market share. Notably, in terms of phones sold, the company has dropped from 18.7% in Q4 of 2014 to 11.8% in Q2 of 2016, showing that they need to do something big to get that 6.9% back.

However, when they were at peak market share in Q4 2011 (23%) they still had a considerable edge on the competition, with companies like HTC and Samsung in catch-up mode at the time. Today, the differences are slim and in many areas these companies are out-innovating Apple. Introducing a better camera, a high definition screen or any other iterative hardware improvement, is not going to do much. Even just launching a phone that looks totally different will do little to put a dent in this decline.

Instead, they may well be looking beyond hardware and to some extent, beyond even the software used. They are looking at their use of machine learning and artificial intelligence.

It is something that many companies are trying at the moment, with Google, Microsoft and Facebook all introducing personal assistants that are equivalents of Apple's Siri, originally created in 2010. Since then, with the introduction of technologies and collaborations with companies like Wolfram Alpha (who also work with several other companies), Siri has improved, but not to the extent that its potential is crying out for. One of the clearest examples of this was their move to a neural net based system in 2014, which significantly improved the virtual assistant, but still didn't allow it to achieve what many expected. In fact, few even noticed a difference.

However, with the impending release of iOS 10 and the inclusion of machine learning within Siri, this may be about to change.

The results of this are, apparently, beyond what even the team who created it expected. Eddy Cue, Apple’s senior vice president of internet software and services claimed in an interview with Steven Levy for Backchannel that, 'This was one of those things where the jump was so significant that you do the test again to make sure that somebody didn’t drop a decimal place.'

There have been rumblings about Apple's use of AI and machine learning due to the increased number of hires they have made in the area, but it is still considerably fewer than their main rivals in the area. The main reason for this, according to Apple, is that they don't hire the people who are already the widely known experts, they hire smart people and train them. This makes sense given their previous hiring patterns, Jony Ive, for instance, had never designed phones or Tablet computers before Apple, but has made some of the most beautiful ever created.

Those who have been working at Apple have apparently been creating an 'AI Brain' within the iPhone, which is around 200mb of information that tracks a huge number of activities that anybody does on their phone, thereby matching, and often outperforming, the increased data input that companies like Microsoft and Google have from being able to input data from their search engines. This data is constantly updated too, with old data being deleted when it's obsolete, hence the 'brain' always remaining at roughly the same size.

In terms of wider practical use, machine learning is going to impact more than just Siri. To some extent it is already used intelligently across the phone, with things like the ability to identify incoming callers from previous emails received or include appointments in your calendar without you inputting it being prime examples. With the release of iOS 10, the wider uses are likely to increase again alongside the considerable improvements for Siri.

Steven Levy seems impressed by the AI and machine learning work taking place at Apple, but the reality of what this work will mean for the iPhone 7 won't become apparent until its release. Even then, given the fundamentals of AI, the full impact this will have on experience won't be felt until even further down the road. However, given that AI is now one of the most hotly contested technological battlefields at the moment, Apple need to come out ahead of the competition to really make a statement.

Bean small

Read next:

City of Chicago: An Analytics-Driven City