Alexa, Amazon’s intelligent personal assistant is getting a speech upgrade according to a recent developer blog post so that she may be able to sound more “human-like”. This includes learning to whisper, varied pitch, emphasis and while not exactly very human-like, she will also be able to bleep out words.
Amazon app developers can use Speech Synthesis Markup Language, or SSML, a standardized markup language that allows for pronunciation control, intonation, timing, and emotion, when coding for their application. This enables the coding of voice related app “skills” which in turn results in an Alexa that has more personality and depth to her speech.
Of course, developers can also use this to make Alexa sound a lot less human or even alien but, as developers have learned from Apple’s Siri, giving digital personal assistants more human-like personality makes it more endearing to the user. Siri often times throws out jokes, even self-deprecating ones which many find charming.
As Alexa-enabled devices also reach more markets, different localized language, dialect and speech versions are necessary and advancing her language skills can contribute greatly to the devices’ adaptability. This is why Amazon also introduced SSML “speechcons” in the UK and Germany to build more creative voice experiences with UK English and German words and phrases. This includes localized pronounciation such as “aluminium” vs “aluminum” for example, as well as region-specific terms such as “Blimey” in the UK and “Donnerwetter” in Germany.
