Life With Talking Machines
Life With Talking Machines


Speech recognition entering mainstream

Smart machines are appearing everywhere. Researchers have struggled for years to make speech recognition and touch controls feasible, and now the payoff is almost here. Apple’s iPhone, corporate phone centers, ATMs, airline kiosks, check out counters, intelligent cars, and vending machines increasingly are all able to carry on conversations with people.

Dragon Naturally, the biggest speech recognition firm, claims its software is 99% accurate without training the system. Cognitive Code has developed a system – Silvia – that uses algorithms to define speech into concepts, rather than words, and places them in a meaningful context. Leslie Spring, the company’s technology officer, says “You can speak to Silvia using whatever phrase you want.” Vlingo sells speech recognition for cell phones, SimulScribe turns voice mail into email, and Yap converts speech into text messaging.


This is now a multi-billion/year industry and growing 15% per year, with experts like James Glass at MIT saying “Speech technology is going to end up everywhere.” Bill Meisel, editor of Speech Strategy, thinks the field is “on the cusp” and Bill Gates said “In 5 years we’ll have more people browsing photos, searching the Web, and organizing their lives using touch and voice technology than typing. It’s one of the big investments Microsoft is making.”


It can be frustrating to spar with talking machines that often don’t understand what you want, but it’s also interesting. I recall speaking with a “virtual robot” named Julie when calling Amtrack. After making train reservations, getting time schedules, and paying her with a credit card, Julie pleasantly asked if I needed anything else? Knowing this was just a software program, I could have simply hung up. Instead, I found myself blubbering sweet nothings to Julie and thanking her for the help. This shows how we love to humanize robots, attributing intelligence, personality, and emotion to them.


TechCast has for years forecast intelligent computers to replace keyboards and “mice” by about 2013.  See Intelligent Interface.