Affective interaction: why feelings matter

Obviously feelings matter to people, there is at least one case study that shows that if you cannot feel emotions, you cannot make decisions, because you have no way of deciding what matters. But why do feelings matter to computers? Or rather, to how computers communicate with people or people communicate with computers?

It’s easy to think of people monitoring your feelings as frightening. For example, there’s one discussion of what would happen if the computer interpreted your voice, so you’d have to say “good morning” to your computer every morning, and your computer would thereby gather a stack of data about your speech and be able to respond to your emotions. Yes, this all does sound a little like the extremely irritating lift in hitch-hikers guide to the galaxy. And in fact, one of my favourite bits of research on computers and emotion is that people prefer talking to a grumpy, irritable computer than to a smooth ingratiating one. There’s a possibility that this is due to the person attempting to get the computer to be less grumpy. In other words, “Go Marvin!, we really do love you”.

But where computers responding to emotion is already important is in hospitals. Robots are used to take people around, and they need to recognise from people’s movements and expressions how important it is to get out of their way. We like that. Well, I do anyway. That’s robots avoiding us when we’re in a hurry. That’s the automated check-out at the supermarket noticing when it is being outrageously irritating and shutting up. That is the ATM that argues with you.

I think there is scope for developing a computer version of a teenager. That you could practice being irritated with and when it kept arguing back to it you could “SWITCH IT OFF”. Smugly.

Leave a Reply