Advertisement

  • News
  • Columns
  • Interviews
  • BW Communities
  • Events
  • BW TV
  • Subscribe to Print
BW Businessworld

Living Digital: Auto Not Correct

With ordinary objects becoming connected and even social, the amount of input wanted from humans is going to go up exponentially

Photo Credit :

1449825614_Lpwtq3_twitter-shutterstock870.jpg

I’ll be flung out of town for a few says,” I text to a business contact who’s suggesting we meet that week. “But we can catch up sin after I get back.” There’s a long period of studied silence during which I almost lose patience with the conversation. “Excuse me?” he asks, finally.

What’s so difficult that this person didn’t understand? I impatiently scan our conversation and realise my keyboard is doing it again, making utter nonsense of what I’m saying to the point of embarrassment. “I’ll be going out of town,” I said, “We can catch up soon after I get back.”

My keyboard has got me into trouble before. There was that time when my editor commanded me to stop sending him obscene messages. I was utterly baffled and not a little put out until I looked through what I’d written. It’s unprintable.

I’m sure everyone has autocorrect horror stories. Those who use Swype like methods and speech recognition for speech to text are not exempt. Mostly with hilarious results, but equally with startling ones. At least I was fairly taken aback when a colleague asked me to “Bring me a couple of Katherines on your way back, please.” Katherines? Where on earth where was I to find a Katherine and why? Turned out he only wanted some ‘kachoris,’ thank God.

Unsurprisingly, entire spaces such as Damn You Autocorrect on the Internet have been devoted to the hilarious results of autocorrect fail —hilarious as long as you’re not the one on the receiving end. Sometimes though, you can be the victim of your own carelessness. Imagine the plight of a person who typed “Can you come to room for a sex?” And then: “Oops. That should have been for a sec.”

There are workarounds to put Autocorrect in its place. Highlighting or underlining the corrected word, for example. But we all know that sometimes we’re in too much of a hurry to even look back at what we wrote — or dictated.

The plot is truly going to thicken as we enter the era of the Internet of Things (or of everything, as Cisco points out) and a time when there are bots and virtual assistants embedded everywhere.

Right now, we mostly get into autocorrect trouble with our phones, but wait till the whole day involves inputting language. Already, if you want a bunch of medicines, you need to type your questions and requirements into a friendly bot waiting at the other end. Or you need to tell an Amazon Echo what tickets to book. Cortana, Microsoft’s virtual assistant with a voice, has to get your work files for you — all you need to do is say. And Siri of course, knows everything so just tell her what you want to post on Facebook, why don’t you.

With ordinary objects becoming connected and even social, the amount of input wanted from humans is going to go up exponentially. And with it, the potential for mistakes. Voice recognition is even more iffy than text input is, and using it quite liberally as I do because I’m too lazy to type everything into the screens around me, I can only say there’s a long way to go before I feel my AutoCorrect really understands me and before I’m convinced it doesn’t have a dirty mind. Maybe someone will come up with an AutoCorrect to correct AutoCorrect. Until then, I’m in acute danger of ordering Katherines using a friendly neighbourhood app.