A few years ago, I was pitching one investor who was put off by the Ubi not having a face. “It needs to be cute,” he said, “people won’t trust it.” This is before the Echo or Google Home. Now, it’s easier to say that an embodied voice is probably going to be creepier. What if you knew what Alexa looked like? Even HAL 9000 had a face as the device. Computer on Star Trek? You couldn’t really associate any one object to it — it was everywhere.
Embodied voices have also fallen flat on their cutesy faces (pardon the pun) in UX with robots in China. Jibo wasn’t a runaway success and Pepper is interesting. However, Alexa, Google Assistant, Siri, Bixby, and others have much more universal appeal because they’re disembodied. We can see them as we wish.
This is the same way books can be enjoyed differently than video and how radio, podcasts, and music can have a profound effect. If we expected to interact with a face, it would draw much more cognitive load. We’d also have to read for emotional cues that wouldn’t work well and we’d be put off in some way.
Better to keep these voice assistants in the ether so we can interact with them better.