To talk or not to talk, that is the question…

When analysing user needs, they all lie within our social space. Our lives are socially regulated within social norms in any given context. It got me to thinking, surely this aspect of human behaviour is the most powerful in UI design. I suspect it can be the most powerful if the object of design has a role in regulating our social space.

In my research I have found, when you talk to a machine, if that machine understands you and reacts to your speech input it becomes a social actor in your life. You apply the same social rules and expectations and treat the machine as you would a human (Reeves, 1996).

I have been thinking of how this theory can help us further understand why VUI just isn’t being taken up by users on a mass scale. The theory suggests that simply talking to a machine or device will automatically enter it into our social world. Humans are made up of many social factors of which speech is part but can speech alone make us human? Probably not. We are much more inclined to be drawn to a voice that has some extra information to back up the voice and aid our understanding. One of the most obvious things that pops into my mind is a face, even when we talk to a person on the phone, for whom we have never met, we are inclined to picture how we expect this person to look from the sound of their voice. Or at least it is a natural expectation that the voice actually has a face.

So, in terms of voice technology and in particular voice synthesis, this brings me to thinking, can we psychologically deal with a voice that has no face? Furthermore, can we humanise a device that speaks to us that we know doesn’t have a face. Let’s here take it back to the mobile phone, there is nothing physically human about a mobile phone, compared to say a car which can be seen to have headlights for eyes, a hood line for a mouth, and turn signals for facial expressions (McCloud, 1993; Norman, 1992). Perhaps that’s why we feel comfortable naming our car. I know I have never named my mobile phone.

The point is, talking to a machine comes with social aspects, unless these social aspects are backed up with more social information, I doubt simply interacting through voice makes a machine human. So what does this mean for our research? Well perhaps the user issue with VUI is rooted in the fact we don’t want to talk to a device and create a social actor unless that device has some sort of extra human information to begin with. It’s not so much ‘I talk to my mobile therefore it is now human’ but more ‘I need to see the human in my phone in order to feel comfortable talking to it’. So is the key to VUI success personification? Perhaps. If we are to accept using a voice interface in our mobile phones we need first to have some personalisaion of that object. If that can be achieved then maybe users will be more willing to talk to their mobiles and enter them into the social space through the use of voice. Or put simply, perhaps that voice needs a face.