Automatic speech recognition (ASR) used to be a concept of science fiction. Increasingly however, it is becoming a staple of our everyday lives. When you call for tech support and are asked to speek the name of the category that you would like to access, you are using speech recognition software.
I was recently sent a great infographic (displayed below) that presents the history of ASR, its modern use and the future outlook for ASR services. The infographic was originally published by MedicalTransition.net.
One of the most interesting sections presented by the infographic is the section on ASR accuracy. In 1993, ASR systems were 10% accurate. By 1995 they were tweaked to be 48% accurate. In 2001 the system were more than 80% accurate.
Based on the pace that the accuracy of ASR systems has advanced, combined with the fact that humans are 96% accurate when listening to speech, it won’t be too long before speech recognition devices are more accurate at transcribing our own language than we are!
There are, of course, some problems when using automated speech recognition systems. Some words like “I” and “eye” have the same spoken sound but are transcribed differently. As ASR technologies advance, I believe we will see them gain the ability to understand the context of a particular sentence to better help them accurately transcribe the proper word or phrase.
ASR products will continue to be perfected and used more heavily. Although typing on a keyboard is relatively easy, it is very difficult to find someone who can type as fast and as accurate as they can talk. Based on the fact that we live in the modern world where time is of the utmost importance, ASR will help humans increase their productivity by allowing us to more easily and more quickly run search queries, compose vital documents and manage our everyday lives.