Everyone is so eager to fire the translators. There seems to be this collective social expectation that those pajama-clad ranks will soon be decimated by a host of friendly artificial intelligences. Picture row upon row of burnished steel automatons typing away in vast warehouses in secure desert locations. Sure, someday. My guess is about five years after everybody else loses their jobs, because translation is like, hard to do.
But I still get suckered. It’s partly the fault of this blog, since I am always reading about these super-duper technologies that are coming along to put me out of business. Latest contender: Lexifone “a robotic interpreter that translates conversation over the phone on a pay-as-you-go basis.” Review time. Gotta check it out.
Sign up. Since the user interface appears to have been localized by non-native speakers with a user interface designed by non-user-interface designers, it took me about 45 minutes to sign up. The first time a user calls in, the tool does a short voice recognition training to help the tool recognize the new subscribers voice. But that’s where this review ends, because I couldn’t understand the instructions of the training tool. I couldn’t even understand my own digitized voice played back to me. So I gave up.
Now, remember, I’m the telephone interpreter guy. The future of my company, my financial independence, my daughter’s college tuition, it all hangs in the balance based on whether or not this technology will put me out of business. Yet I couldn’t complete the training exercise. I couldn’t understand what the machine was saying to me. It’s useless, as audio translations have always been.
The Lexifone site includes three recommendations from people who appear to work as stock-photo models who claim to use the tool for business. This kind of Wall Street ballyhoo directed at investors unfortunately has a big impact for customer expectations too, which is probably where these technologies pose the greatest threat to my business.
I’m convinced that cross-language communication represents the ultimate in user interface problems. I’m also convinced that there is a lot of cognitive bias involved in thinking about foreign language and culture that blindsides a lot of developers and investors, which leads so many astray.
Now I hate technology that costs me money as much as the next guy, but I’m not just throwing around sabots like some Luddite with a computer chip on his shoulder. In fact, I embrace the retirement prospects offered by improved translation technology.
So coming up, we’ll look at a very interesting project that I suspect will set the stage for the next wave of improvement in translation automation. Secret sauce: Deep understanding of how people actually use language. Next time.