comments:
The challenge that excites me is to develop interfaces capable of recognising the kind of emotional colouring that is the norm in face-to-face communication between humans. The challenge is linked to a stock observation about human-computer interaction. At present, when humans communicate with computers, the humans have to make substantial adjustments, and communicate in ways that suit the computers. It is an obvious long-term aspiration for interface technology to reverse the situation, so that computers can – if need be – adopt modes of communication that humans find natural. Natural human-to-human communication includes emotional colouring as an integral part. People continually give each other signs of their emotional status and orientation (to the topic of conversation, to objects or events in their surroundings, and to each other). The process usually goes unnoticed simply because it is so routine. Its absence is more likely to be noticed. People are disturbed by individuals who are consistently emotionless, and professional opinion suggests they are right. Lack of emotional expression is recognised as a warning signal in both psychiatry and counselling. Not all HCI applications would necessarily benefit from sensitivity to everyday emotional colouring. For instance, if a person is using a machine to calculate or keep records or make bookings, the best interface may be one that prompts the person to communicate in a way that is precise and somewhat mechanical rather than spontaneous and natural. However, there are applications where it would be a real advantage if the machine could engage in the implicit transmission of information about emotional states that humans find natural. Examples include: • systems offering advice on subjects with emotional implications (such as large scale choices over finance or careers or education) • artificial sales agents • artificial tutors • artificial companions in computer games. Broadly speaking, if a human carried out these roles, his or her success would depend on responding appropriately to emotion-related signals; and it seems very likely that the same would be true of a machine. Responsiveness may also matter when human beings have reduced ability to adapt their communication to suit computers – because of age, disability, load associated with another task, or whatever – and tend to communicate in the ways that they find natural, even if they are not altogether optimal for the task. As yet, very little technical research has attempted to deal with emotional colouring of the kind that people naturally signal and register in situations such as advice, sales, or tutoring, or companionship. The field began with restricted and often artificial challenges, and movement towards dealing with more natural and diverse types of data has been gradual. Surely one of the big challenges to Humaine is is to get the everyday on the agenda. Roddy
I would suggest senstivity to any or all of the following affective emotional states: Confusion, bewilderment, perplexity, anxiety, frustration, annoyance, fatigue, boredom, distraction, worry, obliviousness, apprehensiveness, embarrassment.

AISB 2009 Symposium on Affective Bodily Expression
