Here’s why Microsoft bought SwiftKey for $250 million—artificial intelligence


Photo: BloombergPhoto: Bloomberg

New Delhi: On Tuesday, London-based predictive technology company SwiftKey said it has joined the Microsoft family. The deal, first reported by the Financial Times, is believed to be around $250 million.

SwiftKey, for most smartphone users, comes through as just another third-party keyboard app, an upgrade on the default keyboard that most platforms offer—easier to swipe and therefore, easier to type. But beyond the app and its user interface, what the company offers, and something that makes it attractive for the likes of Microsoft, is the very technology that powers the app—artificial intelligence or AI. In SwiftKey’s case, its predictive technology learns from the user’s personal writing style, on the basis of previous usage of text, chat messages or social media interactions, and predicts the next word or phrases that the user intends to type. This, as the company on its website says, could be anything from “your quirkiest family name to the sports teams you support”.

The company, which was launched in 2008, says its “mission is to enhance the interaction between people and technology,” with the fundamental belief that “technology should adapt to you and not the other way around.” And therefore, it came up with what it calls the “world’s most intuitive and personalized keyboard software.”

Harry Shum, Microsoft’s executive vice-president, technology and research announced its acquisition of SwiftKey in a blog post. He wrote, “This acquisition is a great example of Microsoft’s commitment to bringing its software and services to all platform. We’ll continue to develop SwiftKey’s market-leading keyboard apps for Android and iOS as well as explore scenarios for the integration of the core technology across the breadth of our product and services portfolio. Moreover, SwiftKey’s predictive technology aligns with Microsoft’s investments and ambition to develop intelligent systems that can work on the user’s behalf and under their control.” Put simply, Microsoft has just entered the already crowded (and growing) artificial intelligence game.

It is also following in the footsteps off fellow tech giants, most notably Google and Apple. Google, in January 2014, paid $400 million to acquire another London-based AI start-up, DeepMind. The deal went down as Google’s largest European acquisition to date. DeepMind, according to aTechCrunch report, was founded by “neuroscientist Demis Hassabis, a former child prodigy in chess, Shane Legg and Mustafa Suleyman.” Besides, the company also had Skype and Kazaa developer Jaan Tallin as an investor, prior to Google’s acquisition. Prior to acquiring DeepMind, Google, in December 2012, had appointed “famed inventor, entrepreneur, author and futurist” Ray Kurzweil as the director of engineering, focused on “machine learning and language processing.”

Similarly, Apple made its first acquisition of an artificial intelligence outfit in October last year, when it purchased Vocal IQ, another startup, based in Cambridge, UK. Vocal IQ, according to a Forbes report, specialises in “speech-related artificial intelligence.” At the team, Apple’s acquisition was expectedly linked to the company wanting to enhance its virtual assistant Siri’s capabilities, but Forbes reported that Vocal IQ, in 2014, worked with carmaker General Motors “to integrate its capabilities into an intelligent voice-controlled system for its cars that would let users turn on their windshield wipers or adjust car stereo settings by speaking.”

And then, there’s Facebook, which in December 2013 announced that one of the world’s leading “deep learning and machine learning scientists”, Yann LeCunn, who was a professor at NYU. LeCunn was hired to lead Facebook’s new AI laboratory. Facebook, today, has what it calls FAIR (Facebook Artificial Intelligence Research) based in New York, with other offices in its Menlo Park campus, London and Paris. The company says that it is “committed to advancing the field of machine intelligence and developing technologies that give people better ways to communicate. In the long term, we seek to understand intelligence and make intelligent machines.” And as if to lay down a marker, Facebook said it hopes to accomplish all of this by “building the best AI lab in the world.” You can read more about Facebook’s AI efforts here.

This brings us to IBM’s Watson, what it calls a “cognitive technology platform that uses natural language processing and machine learning to reveal insights from large amounts of unstructured data.” The system, in the case of Watson, essentially learns everything (the subject concerned), with pairs of questions and answers related to the subject. When it comes to answering (a question related to the subject), it scours through “millions of documents to find thousands of possible answers.” It also collects evidence and “uses a scoring algorithm to rate the quality of this evidence,” before ranking “all possible answers based on the score of its supporting evidence,” all of this according to its official website.