Deep learning, the idea that a machine will think for itself when given access to data, rather than having to be fed information, enables a machine to learn from its mistakes and grow. It’s not a new concept, but in recent years, some of the world’s biggest companies, including Google and Apple, have begun to exploit the technology to improve their operations. Elements of deep learning are already being used by Apple, with its personal assistant Siri learning from your actions and questions to develop an idea of who you are and what you like. Further afield, deep learning could be used for things like medicine and healthcare, potentially saving billions of dollars on expensive treatments and ensuring that patients are cared for correctly
DEEP LEARNING: BIG BROTHER IS WATCHING YOUR EVERY MOVE
Although there are some significant benefits to deep learning, it does bring with it privacy concerns, with many suggesting that there is little limit on what companies such as Apple and Google can find out about their users. Google, in particular, keeps hold of an unlimited amount of data from its users, with the company even scanning boarding passes to give users reminders, with seemingly little warning or permission.
It seems everything we do online can be traced back to us. Search for some shoes you want on Google, an advert for them will appear on Facebook. It’s the new world we live in, and it’s part of deep learning and artificial intelligence – but just how far will it go?
HOW FAR WILL IT GO?
Could deep learning be the beginning of something more sinister? The movie HER (2013) explores the possibilities of artificial intelligence and follows the story of a lonely man in the middle of a divorce, who purchases the new OS1, the first artificially intelligent operating system. However, this system is more than just a computer chip – it’s supposedly conscious. The film explores the relationship between machine and man and the extent to which machine consciousness is possible, and is the first signs that deep learning and artificial intelligence is taking over. Fast Company explored the extent one individual went to escape online tracking. To do so, he had to create an entirely fabricated, new identity. He did not intend to assume this identity but experimented the length to which you must go to have an untraceable identity. Curtis Wallen entirely created Aaron Brown for what he refers to as an art project. He used the cryptocurrency Bitcoin, as this does not need to be linked to a real person in the way that PayPal does. He finally resorted to creating a new identity each time he contacted a new person. The lengths he had to go to prove just how hard it is to escape online tracking and demonstrated that nobody can escape the data collection from companies such as Google, Amazon, and Apple.
WHAT IS DIFFERENTIAL PRIVACY TECHNOLOGY?
In recent times, Apple has recognized the frightening reality of deep learning and is now striving to protect its users. With the release of iOS 10 last year, Apple achieved the seemingly impossible, finding a way to use customer data without compromising on personal security. By utilizing differential privacy technology on its iOS, macOS, watchOS, and tvOS operating systems, the company aims to keep data “completely private,” writing in a press release: “We believe you should have great features and great privacy. Differential privacy is a research topic in the areas of statistics and data analytics that uses hashing, subsampling, and noise injection to enable…crowdsourced learning while keeping the data of individual users completely private.” In simpler terms, differential privacy is essentially a system that allows Apple to learn as much as it can about the data its users provide while learning as little as possible about the individuals who produce this data. Essentially, personal information is stored on personal devices rather than the wider Apple infrastructure. The technology is mathematically proven, which does a lot more than just remove names from data. It’s been described as a futureproofed system by some. iOS 10 was the perfect time for Apple to implement this new system, as the new release also brought with it predictive software such as QuickType and a more advanced spotlight search where personal information like contact details, emails, and photographs were to be used. For companies like Apple, it is not considered an option not to analyze and utilize their user’s data, as it is crucial for the expansion of their company and for producing effective and user-friendly software. Instead, the Cupertino firm has created a solution that protects the privacy of their customers, while allowing them to exploit this data on a deviceto-device basis for the benefit of the end user.
DEEP LEARNING IN THE WIDER TECHNOLOGY WORLD
Apple is not alone in using deep learning and providing innovative solutions that will protect the data of their customers, however. Some of the world’s biggest technology companies, such as Uber and Microsoft, are beginning to see the advantages and are using the technology to power their operating systems and smartphone apps. Uber has ensured that they are not exploiting customer data while making sure they can offer the best possible services. Similarly, Microsoft has said that they will make their data available to researchers and government agencies while ensuring that personal information cannot be traced back to individuals.
TOO GOOD TO BE TRUE?
Although Apple’s solution to deep learning and data protection has been praised by the majority of technology critics and security analysts, Matthew Green, cryptography professor at John Hopkins University, was not convinced. “Most people go from theory to practice, then to widespread deployment. With Differential Privacy it seems Apple cut out the middle step,” he tweeted. Matthew went on to say that he believed that Apple’s data protection software was a “nice idea” but “it ends up being a tradeoff between accuracy of the data you are collecting and privacy.” He argued that the accuracy decreases as privacy increases. In his eyes, differential privacy is relatively untested and even possibly dangerous. And Green is not alone, with others are beginning to take issue with the technology. An article by iDownloadBlog states that: “Differential Privacy can be a double-edged sword and many folks were quick to point out that Apple’s refusal to collect huge amounts of data on users, like Google is doing, is hurting its ability to compete in the AI space.” Only time will tell exactly how beneficial differential privacy can be.
THE START OF SOMETHING BIG
In the face of privacy issues, cryptocurrency is on the rise. Originally, Apple was entirely skeptical about cryptocurrency and still has an extensive approval process for iOS applications. If Apple does not deem a cryptocurrency fit, then it does not grant app developers with the means to develop apps around it. Apple does now accept Zcash, and it is now available for iOS. At WWDC Apple announced a new feature for iOS 11 which allows users to easily send and receive money via iMessage, free of charge. However, this is thought to be the start of something bigger. The Apple Wallet will soon feature a new card called Apple PayCash. According to an article by The Mission: “Some people thought of it as a gift card because you can use this money to pay through ApplePay in a cab, at the grocery store, or on the web.” However, The Mission states that they believe this is, in fact, the beginning of Apple’s own cryptocurrency. This is because it is described as a digital currency that stores value, which is exactly how people describe Bitcoin, the first and most famous cryptocurrency. According to The Mission’s article, people are afraid to trust cryptocurrency because you just can’t trust them in the same way you can your bank. However, “with AppleCash, you would gain the benefit of an increasing user base, Apple’s stock price, and continued R&D, the cash and real assets on their balance sheet,” so only time will tell whether the currency will be adopted by users.
APPLE REMAINS ON THE RISE
Despite doubts over Apple’s deep learning and differential privacy solutions, there’s little doubt that it is a company on the rise. With the release of iPhone 8 this September alongside iOS 11 and an array of software releases for watchOS, macOS, and tvOS, it’s believed Apple will become a trillion-dollar company within the next twelve months. According to Wall Street analyst, Brian White, Apple has “the most underappreciated stocks in the world,”, and it’s clear that, as the firm looks at new ways of keeping customers within its ecosystem (this time with a new cryptocurrency), the giant will continue to dominate the technology market and keep customers, and dollars, on its side.