Experiences with Speech Technology

Many technologies are introduced well before they're ready for regular use. It also often takes a combination of factors to come together to enable the successful adoption of a technology. The alternative interaction technologies - touch, gesture, and speech - have all gone through this transition. They started off in research labs with demonstration experiments shown regularly at academic conferences. Then the technologies moved into product development labs and eventually into early products. Touch is now well establshed pervasively across many markets by being very successfully used in smartphones and tablets. Gesture was introduced in the casual game market with consoles. It represents a novel and interesting interaction technology but doesn't appear to be compelling enough to be heavily used even in that market although it does still show significant promise. Speech technology has been around the longest in some commercial form, probably has the greatest promise, and yet it has been the one that has proven to be the toughest one to move into widespread adoption and use.

Speech dictation especially in specialized domains has been successfully used for many years. However, it was the development of smartphones, high bandwidth connections, and cloud technologies that provided the environment for speech technologies to be adopted more widely and generally. I now use speech technology to select the person to dial, execute some searches, and for some dictation. I find that it works quite well but I'd like to use it more pervasively than I do now. I also wondered what other people's experiences have been with speech technology so, as usual, I turned to the social networks and asked "Do you use speech technology? If so, how and for what? If not, why not?"

A total of 53% of those who responded said that they used speech technology.  The themes regarding use of the technology stressed hands-free contexts like driving, riding a bike, running, as well as simply "laziness and convenience". Most appeared to be quite happy with it.  Those who didn't use the technology gave the following reasons why they didn't.

  • Slower than a keyboard
  • Too error-prone
  • Too easy for others to listen in
  • Not good in noisy environments
  • Too much effort
  • Not accurate enough
  • Too inconvenient
  • Doesn't work everywhere that a keyboard does
  • Doesn't recognize Indian accents very well
  • Doesn't recognize French accents very well
  • Have to repeat too often
  • Needs to be more readily accessible

It would appear, therefore, that speech technology is coming of age and successfully used by some people quite regularly typically when they can't or don't want to use their hands. However, the technology still has to improve in a number of ways including speed, accuracy, pervasiveness, and globalization to be used by an even greater number of people. I think we're on the cusp of the effective pervasive use of speech as an important interaction technology.