This technology can type three times faster than humans—and could be a tech job creator
Speech recognition software just ramped up in a big way—and so have opportunities in this growing field.
As much as we love to make fun of Siri, speech recognition software types faster than humans — three times as fast. A new study by Stanford University, the University of Washington and Chinese search giant Baidu pitted cutting-edge speech recognition software against human texters using an iPhone keyboard, and even those with exceptionally fast thumbs couldn’t compete with technology.
The subjects took turns typing simple spoken phrases like “wear a crown with many jewels” or “do you like to go camping” while an app recorded their typing speeds. Half of the human participants were native English speakers using a QWERTY keyboard, while the other half were native Mandarin speakers using iOS’ Pinyin keyboard. In both languages, the artificial intelligence was faster: three times faster in English, 2.8 in Mandarin. The software also had a lower error rate than human typists in both languages.
Long story short? The time is right to get in on the speech recognition job market, and there are already plenty of opportunities out there: Amazon is looking for a data engineer for its cloud-based voice service, Alexa. AVA is seeking a Senior Data Scientist who specializes in speech. Verizon is looking for a voice technology project manager to help bridge the divide between touch tone and speech-based call services.
It’s a high-growth area: More and more people use voice assistants, whether it’s to find information or write out a hands-free text. In fact, a 2016 Internet Trends Report found that 65% of American smartphone owners use the technology.
That’s a number that could have some surprising influence on the job market. Earlier this year, Monster reported on an Apple expansion that will open Siri up to outside developers—and as major apps work to integrate Siri, jobs working with the platform could get a serious boost.
New developments in speech recognition and AI technology could heighten fear in workers who are afraid of getting replaced by robots, but here’s some good news: The Stanford study didn’t indicate how well the speech recognition software would perform typing full paragraphs, or its ability to construct sentences with proper punctuation. So at least humans are still good for something.
Of course, you could always become one of the people building the robots.