There is increasing interaction between humans and avatars, for example in video games, and between avatars- that embody interacting humans in on-line communities such as Second Life.
Artificial and intelligent agents might take both human and non-human forms. They might be simply virtual constructs or engineered robots. Both have advantages and disadvantages in their visual appearance and whether they are constructed in pixels or machine components.
There are practical, psychological and aesthetic considerations in the embodiment of artificial agents in both how they are displayed and whether they cause any unease in what they look like and how they behave and in what kinds of interfaces are possible with them.
There would not be an issue of scale for an on-line avatar. But an intelligent agent with a robot body might be disconcerting and difficult to communicate and interact with if it is too small or too large. Our unease with small insects and large animals point to the problem we might have with non-human robots- even if we can share behaviours that might indicate such expressions of surprise, anger, fear, compliance and contentment.
An advantage with robotic embodiment would be that the human and the intelligent agent would be able to interact in 3D space, with but with immediate physical, real-world consequences. Interaction in a virtual space can be simulated seductively, but with consequences only within the virtual task environment.
With the increasing possibilities of sensing and vision systems, the user will be able to interface, actuated and communicate with the avatar with hands free to gesture rather with hands-on keyboard or game joystick devices. And given the increasing proliferation of haptic technologies, the user will not only see the avatar but be able to experience both touch and twist (texture and force-feedback) through either proximal or remote interaction.
Authors: stelarc
Event: SF08: Embodied Interaction in Mobile, Physical and Virtual Environments Workshop