Engineers nowadays have to think about an important feature on any new piece of technology: Accessibility. This can be done differently in order to adapt to specific needs. Say, a person wants to use a smartphone but he or she has a vision impairment, the company that made the phone can’t lose a possible customer, so they have to take this barrier into account and find a workaround.
After some adjustments, the device is able to read what’s on screen to the user and can be controlled by voice commands. There are many ways technology can be made accessible to the public that has difficulties accessing them.
My Intention in this Extended Essay is to talk about a new technology that is tangent to the topic before and has been improving each day with more research into Artificial Intelligence: Voice activated technology/Personal assistants, while analysing their impact on the disabled public.The development of voice recognizing machines is often compared to a child growing up. it slowly learns vowel for vowel, then word for word, phrase for phrase until its state nowadays, where it can comprehend difficult compositions of terms and reply with an incisive and coherent answer. Firstly, let’s analyse how the main component of these assistants was created.Voice recognition technology, is something present in all personal computers and smartphones since 2011 and it has become so normal to be able to talk to machines that we’ve taken it for granted, although, its history goes way back to the end of the 1700s, with Wolfgang von Kempelen’s Acoustic Speech Machine.
This machine was something called a speech synthesizer and all it could do was make sounds that were similar to the vowels “O” and “A” and the consonant “M”. It was after this invention that the workings of human speech became more studied and the study of phonetics began to become more popular. It was first needed for humans to understand speech before machines could replicate it. Later, in 1879 Thomas Edison invents the dictation machine, an apparatus that is able to record sounds into a physical form, and enables it to be played back again using the same recording.Succeeding this leap in technology, was the invention of computers, which opened the doors for Audrey a system by Bell Labs.
Created in 1952, Audrey was a computer that could recognize digits with a 90% accuracy rate, but it only worked with it’s creator’s voice, which lead it back to the drawing board. 10 Years later, in 1962, IBM released its so called Shoebox, a computer not only capable of distinguishing between 16 english words, but also acting upon those said words. Who would have thought that in less than a decade from Shoebox, scientists at Carnegie Mellon University would be able to outperform IBM? In 1971, Harpy was born. A system that broke all records at the time with its incredible 1,011 words database, it could also understand simple phrases. In spite of such advancements, this technology wasn’t even close to being publicly available. It could only be found in laboratories and research centers, which at that time were funded by governmental organizations, like the Defense Advanced Research Projects Agency (DARPA) in the U.S.In 1986, IBM’s Tangora was the new system in the scene, it made use of a Statistical Markov Model called Hidden Markov Model (HMM), which enabled the machine to predict the following phonemes in speech to assist the machine’s abilities at comprehending the user. From this point, the once primitive technology, could not only hear people but also effectively listen to them, being able to make its way up the business world and the medical workspaces as a commercially available product. While it was functional, it wasn’t very practical as it couldn’t fully understand a person’s normal speech flow, meaning that a pause had to be made in between of every word