<div><em>Imagine what it would be like if your phone more or less auto-completes an email for you. Or makes quick work of a text message when you’re rushing about with little time on your hands. <strong>Mala Bhargava</strong> explains</em></div><div> </div><div>That’s what the future of keyboards will be like and the first off the mark is Swiftkey, a predictive keyboard app that is already used by millions. Other than the usual Swiftkey Keyboard app that you’ll find on Google Play, a test version of another Swiftkey app is out, called Swiftkey Neural. You can get to it by typing the name into Google’s search and picking it up from the Swiftkey website. But it’s still in Alpha, a very early stage of testing, and not really meant for publicconsumption just yet though those able to test it out are welcome to download the app and select the new keyboard. It’s still also limited and has only one language – US English.</div><div> </div><div>Swiftkey Neural uses artificial neural networks to predict and correct as you input text. The “Flow” mode is also there in case you like to use the tracing method of entering text. The keyboard may not look or seem different to anyone using it at first. But it’s long-term usage which will pick up on the phrases and language patterns you use typically and use that in text prediction, making it startlingly accurate. The keyboard and the way it works is based on a field of artificial intelligence or AI which tries to replicate the way the human brain works. Michael Smith, VP of Consumer Products at Swiftkey says that language, after all, is all about patterns and it’s patterns of expressions that are going on in your head. And it’s exactly this that is understood by artificial intelligence.</div><div> </div><div>It’s beyond impressive that technology that can understand context and the actual meaning of a phrase or sentence can now be housed in a relatively small device that fits in your pocket. At one time, one needed computers that filled a room to achieve the same thing though of course it wasn’t even the same thing.</div><div> </div><div>The convenience that a fast accurate comprehensive predictive technology is going to bring to the phones we all carry can’t be underestimated. People hate inputting on glass, and quite rightly too, and the fact that entering text will be just a matter of a few taps isexciting. Those who have been hanging on to barely-working old phones with physical keyboards and not moving up to touchscreens because they think they can’t get used to it will no doubt find it much easier with predictive text that’s practically on steroids. But of course, with getting something that clever, you also give up something and that’s handing over a little more control to machines. Already Google guesses what we’re searching for, Facebook shows us what it thinks we want to see, and so on. The extent to which technology is going to predict stuff for us and what we’re going to do with our brains while all that is going on somewhere in someone else’s server is a bit of a worrying thought. This is specially so when you consider that once the technology is perfected, in use, and miniaturised, it will find its way into more machines, including those that use voice recognition. It isn’t inconceivable to think that there will also be predictions based on a mere word or tone, not too far in the future. The fun has only just begun.</div>
