Apple Incorporates AI Functionalities Without Compromising the Privacy

The potentials of AI are so influential that it has made every tech giant go frenzied about it. Google, Microsoft, Amazon, Baidu…and so on, you name it and they all have their secret sauce for AI success. Several acquisitions, mergers, partnerships, investments, and funding prove that they are quite serious about their innovation in this space and attaining the title of the lead innovator. Even home appliances and electronics-fame Samsung has released its AI-human, Neon, at CES this year. The company that gains thunderous buzz every year with a new launch of its iPhone series, Apple has been lagging in this race for a while. However, the company too has now become serious in battling the AI race with the aforementioned companies.

Recently, Apple acquired Xnor.ai, a company that creates AI software for mobile devices, for a reported US$ 200 million. It is anticipated that the acquisition will apparently lead to AI applications running directly on iPhones, iPads, etc., not outsourced to the cloud, which would make these tasks more private and quicker.

Xnor.ai explained its further goals as "We're unleashing the power of AI to run as efficiently on small inexpensive devices as it does in massive data centers. And internet connectivity isn't necessary anymore. We call this AI everywhere." The startup's technology could potentially be used in Apple's Siri voice assistant. As of now, most spoken commands are uploaded to servers where AI is used to transform the sounds into words an iPhone can understand. Apple aims to conduct this process on the mobile device instead, to make it much more private so that only the iPhone would ever hear what users say.

As noted by an article, "while there's no shortage of startups in the current AI boom doing machine learning (ML), Xnor.ai's ambition is a bit different. Whereas everyone else is combing through massive amounts of data to create smart tools, Xnor.ai focuses on building AI algorithms which run locally on devices, rather than in remote data centers." This subsequently means availing the perks of AI without any odds and negatives.

This isn't the first time Apple has shown this balancing act commitment in terms of its privacy when it comes to AI. The company had acquired Silk Labs, an AI startup that does image and audio recognition for people detection, facial recognition, and more, in late 2018. There too, all the processes are carried out locally, without sending data to the cloud. As per Silk's now-defunct website, "privacy and security are built into our company's DNA. With every line of code, we write and in every design decision we make, Silk takes great measures to ensure that user data on the Silk Intelligence Platform is fully protected at all times."

Moreover, Apple has also begun to share significant details about its privacy-focused AI prowess. The company published a paper last year, describing something called federated learning. Notably, federated learning is a technique through which machine learning algorithms can be trained on multiple local datasets without exchanging data samples.

Through this Apple, can get Siri to recognize users' voices (and only their voice) as a wake word. However, it does it without sharing this data with a data center. The samples which are shared are the updated neural networks, which are used for improving the overall master network.