View the slides for this talk here.

Over the last five years we have been investigating new forms of human-computer interaction (HCI) based around the concept of 'Companions'. In the context of a three year project looking at technologies for older people we explored the ideas of artifical companions that could move across devices and that were 'personality rich'. A two-week summer school explored the idea of a Companion that accompanied people to a large arts festival. Currently we are exploring the ideas through a four year EC-funded project called Companions. We have also referred to this concept as 'personification technologies' (Benyon and Mival, 2008; Mival, Cringean and Benyon, 2004) because the aim of these new forms of interaction is to encourage people to personify the technology and to attribute human-like characteristics to it. Wilks (2007) has also characterised the Companion concept. He sees it an intelligent, personalised, persistent, multimodal interface to the Internet. Drawing upon recent advances in Human Language Technology (HLT), inference and knowledge representation, he sees a Companion as a humane conversational partner.

Designing for new and possibly disruptive technologies represents a particular challenge for interaction design. The usual process of interaction design consists of requirements generation, conceptual design, physical design, prototyping and evaluation (Benyon, Turner and Turner, 2005). Requirements generation typically concerns undertaking field work, observing people using existing systems and asking people what they would like. With the wholly new experiences that are expected of Companion technology, such approaches are often inadequate. People find it very difficult to imagine and comment on experiences until they are engaged with some concrete examples of those experiences. In this paper we discuss how to envision such experiences using 'wizard-of-Oz' techniques, mock-ups and video scenarios.

Companions will make use of novel speech and language technology and techniques to enable intelligent and personalised interactions. They will be able to exploit modern input and output technologies such as touch screens and a much more powerful Internet. With these developments we see a significant change in the opportunities for interaction. More intuitive and personalised interfaces using gesture, movement and natural language are rapidly becoming available. This provides the opportunity to move from current graphical user interfaces and a tool-based view of HCI to more engaging interactions with a subsequent greater commitment and involvement of people.

Our view of Companions is that we want to change Human-Computer Interaction into Human-Companion Relationships. This builds upon the ideas of affective computing (Picard, 1997) and designing for emotional involvement. Companions may be represented as a 'virtual human' on-screen character or as a embodied conversational agent (ECA), but they do not have to be. Whilst the term 'Companion' is meant to invoke personification, we see Companions as encompassing the widest possible range of devices and forms of interaction that woven together produce a relationship-building experience for people. A home embedded with ambient intelligence could be a Companion. A character that moves across devices and domains, and understands its owner's needs and wishes could be a Companion.

The grand challenge of Companions is to develop technologies that people have relationships with. This is rather different from the social presence of two people feeling close even though they are distant. It is to do with people forming relationships with technologies, feeling present with a technically constructed entity.

We do not believe that Companions will be manifest as a simple device. Companions represent the next generation of people's interactions with information and with each other. If they are to be successful Companions will require new services and service providers. They will have a lifetime, and perhaps an afterlife.

Benyon, D. and Mival, O. (2008) Landscaping Personification Technologies. In Proceedings of CHI 2008, Florence, Italy

Wilks, Y (2006) Artificial Companions as a new kind of interface to the future internet. Oxford Internet Institute, Research Report 13, October 2006