Alexa, Echo, Google Home and other speech assistants were identified by CNN as a trend to watch for 2018.
However, technology including ‘speech recognition devices’ is not a trend or new idea to individuals with disabilities. In fact, technologies like this have been helping individuals with disabilities for many years. Technology which has been providing accessibility to individuals with disabilities is now being offered in the commercial market as a mainstream accessory and standard feature in our cell phones and smart homes. Microsoft’s announcement to offer technology accessibility grants, labeled AI for Accessibility, is one of the most recent examples of the technology industry highlighting and promoting accessibility features of their devices.
Technology companies are bringing assistive technology features such as voice dictation into the mainstream with devices such as Alexa, Echo, and Google Home. Speech recognition and speech predictability software have been constantly evolving since the 1950’s including the introduction of Dragon Naturally Speaking to the commercial market in the 1990’s and Google’s introduction of Siri in 2010. Alexa is one of the recent iterations of voice activated software and has become commonplace such that St. Louis University recently announced that this year it will be providing all dorm rooms and apartments on campus with an Alexa.
The history of speech recognition and dictation technology began in the 1950’s as systems were created that initially were able to recognize only spoken numbers. Engineers of Bell Laboratories designed the “Audrey” system in 1952 which recognized digits spoken by a single voice. In 1962 at the World’s Fair, IBM debuted its “Shoebox” machine, which could understand 16 words spoken in English.
The 1970’s brought significant milestones in speech recognition technology primarily due to funding from the U.S. Department of Defense DARPA Speech Understanding Research (SUR) program, from 1971 to 1976.
A decade late, the 1980’s, brought a shift from speech recognition to speech prediction. A statistical method known as the hidden Markov model (HMM) allowed technology to not only identify word sounds and templates, but identify possible words and word sounds. Technology grew from being able to identify a few hundred words to recognizing thousands of words. In the 1990’s, speech recognition software became commercially available for the first time in programs such as Dragon Naturally Speaking.
Around 2010, Google changed the landscape of available speech recognition technology by incorporating voice search, Siri, onto its Android phones and also to its Google Chrome browser in 2011. Their voice search capability can identify an amazing 230 billion words!
Siri, Alexa and Google Home are the latest versions of voice recognition devices but with voices that include “added personality.” Most voice operated devices now are pre-programmed to be informative and also to provide funny and witty responses to users’ questions. Technology interaction is now becoming a two-way conversation; as Melanie Pinola says, “Speech recognition has gone from utility to entertainment.”
If we want to predict technology trends of the next decade, we should examine the field of assistive technology and watch for devices/tools currently being used by individuals with disabilities to enter the mainstream market.