This is Lecture 25 of our series on prosody. In recent years spoken dialog systems, smart speakers, intelligent agents, and their cousins have seen wide acceptance, but only in certain domains. Going component by component, we explain how better use of prosody may dramatically improve the usability of spoken dialog systems. Yet future systems, with super-human prosodic abilities, will pose societal risks.
00:00 Towards Truly Interactive Dialog Systems
01:06 Limitations of Current Systems
03:15 Typical Architecture of a Dialog System
04:03 Speech Synthesis and Speech Recognition, revisited
04:49 Turn-Taking and other Reactive Behaviors
05:40 Sensitively Tracking the User's State
07:14 Clearly Conveying the System's State
08:56 Challenges in Scaling Up
11:03 Societal Risks
