Amazon makes Alexa more expressive with new 'skills,' assistant to be 'more human'
Users of Amazon's artificial intelligence-driven smart speaker may soon be able to hear an entirely different sounding Alexa in the coming days. Amazon has announced that they will be implementing a number of new speaking skills for the virtual assistant, which are aimed at making her sound more human.
Some of the changes that will be implemented will allow Alexa to whisper, pause for emphasis, and adjust the pitch and volume of her voice for specific words and phrases. Developers may also program Alexa to "bleep" out certain offensive words to make the device more child-friendly. App developers for Alexa will reportedly be getting access to a brand new type of standardized markup language, called Speech Synthesis Markup Language (SSML), which can be used to alter Alexa's voice in certain situations.
There are of course an unlimited amount of things developers can use the new markup for that may include new "skills" apps for Alexa. According to Tech Crunch, Alexa already has a pleasing personality, to begin with, but the virtual assistant can sound quite monotonous and robotic at times, especially when she is reading long lines of text and neglects to take pauses and breaks, which is unnatural for humans. Amazon wants to change this by allowing developers to have Alexa pause or break at appropriate points as well as change her tone depending on the context of what she is reading out.
In addition, Amazon has also implemented new "speechcons" to the virtual assistant, which are basically words that need to be expressed in a more colorful way as per a report from GSM Arena. Alexa pronounces and says these words differently, which is of course aimed at making her sound more engaging and natural. Words such as "ahem," "eureka," "wow," "yay," abracadabra," and other terms are automatically pronounced differently. Amazon is also expanding this to include regionally specific expressions such as "Blimey," "Bob's your uncle," "Donnerwetter," and other language-specific expressions.