A Vaklempt Alexa

You’re either excited or disappointed.

It was exciting to see Amazon’s announcement that it’s expanded Alexa’s capabilities of speech synthesis with different levels of “excited” or “disappointed”. This adds to the announcer voice it had uh… announced… maybe half a year ago.

This is a continuation of Alexa’s support of speech synthesis markup language (SSML). Already, as a Skill developer, you could get Alexa to announce certain canned responses as a pre-recorded exclamation (e.g. “cowabunga!”). It also allowed for changes in prosody.

SSML support, however, requires some effort on the part of the developer. Beyond just thinking about the text of a response, they now have to think of the expression. This means another 2–3 revisions of material for Skills development.

However, that might not always be the case. What if Amazon were to apply sentiment analysis to the text to be able to automatically adjust the SSML? What if it were to map human prosody for similar language? That could create an automated emotion-laden speech response from any Skill.

We’ll likely see this over the coming years. The first step will be more types of emotion and then next after that, automatic application of emotion from the sentiment of the text or context. It’s going to be an interesting ride!