Voice Recognition Software Finally Beats Humans At Typing, Study Finds : All Tech Considered In a face-off between voice entry and typing on a mobile device, voice recognition software performed significantly better. The results held true in both English and Mandarin Chinese.
NPR logo

Voice Recognition Software Finally Beats Humans At Typing, Study Finds

  • Download
  • <iframe src="https://www.npr.org/player/embed/491156218/491242758" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript
Voice Recognition Software Finally Beats Humans At Typing, Study Finds

Voice Recognition Software Finally Beats Humans At Typing, Study Finds

  • Download
  • <iframe src="https://www.npr.org/player/embed/491156218/491242758" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

KELLY MCEVERS, HOST:

Computers have already beaten us at chess, "Jeopardy!" and Go, and humans have now lost another battle over texting. A new study shows that software is significantly better than we are at typing text messages. Here's NPR's Aarti Shahani.

AARTI SHAHANI, BYLINE: The study is by Stanford University, the University of Washington and Baidu, the Chinese internet giant. Baidu chief scientist Andrew Ng says this should not feel like defeat.

ANDREW NG: Humanity was never designed to communicate by using our fingers to poke at a tiny little keyboard on a mobile phone.

SHAHANI: And testing shows there is a better alternative - talking. Researchers set up a competition, pitting a cutting-edge Baidu program called Deep Speech 2 against 32 humans ages 19 to 32. The humans would say and then type short phrases into an iPhone like wear a crown with many jewels, and this person is a disaster. They found the voice recognition software was three times faster which Stanford computer scientist James Landay did not expect.

JAMES LANDAY: The surprise for me was that it was that much better - three times faster. You would think everyone'd be flocking to use it if they knew how much better it actually was.

SHAHANI: Just like smartphone cameras have more megapixels to see us clearly, the built-in microphone can hear us more clearly. Also, supercomputers have more voice recordings to vacuum in and analyze. Still, voice recognition gets a bad rap. That could be because of how people use it. Many people ask Apple's Siri a basic question and too often get a wacky response in turn.

LANDAY: People probably play with Siri and find, oh, it didn't give them the right answer so they don't think to use speech as a way to do their text messaging or to do their email or whatnot.

SHAHANI: The researchers didn't test query skills. They zoomed in on the ability to spit back the right words in two languages. In English, they found the software's error rate was 20 percent lower than humans typing on a keyboard. And in Mandarin Chinese, it was 63 percent lower. Landay hopes these findings encourage people to talk to their phones more in order to transcribe and text.

LANDAY: Using speech for those things is now working really well.

SHAHANI: It's easy to see how talking at your device would be far better than typing - say, when you're driving. Baidu's Ng imagines another scenario. He does not have children yet, but he says he looks forward to the day when his future grandchild asks him...

NG: Is it really true that when you were young, if you came home and you said something to your microwave oven, would it really just sit there and ignore you? That's just so rude of the microwave.

SHAHANI: His co-author Landay reins him back and notes there are many moments - in a meeting, in bed with your partner sleeping - when typing still makes more sense than talking to one's device. Aarti Shahani, NPR News, San Francisco.

Copyright © 2016 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.

NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.