These are some of our objectives:

• To create novel artificial agents that have sophisticated behaviours, yet have low-fidelity embodiments.
• To instill in the viewer a sense of conversing with artificial agents.
• To develop interactant-representative (IR) agents that are representative of the human interactant.
• To build a hardware infrastructure shaped by the artistic vision of the work, one that extends and augments the infrastructure developed for the artworks Swell [2003] and Flo’nGlo [2005-2006]: abstract in shape and form, using mixed materials, controlled by an underlying computational platform, and incorporating the use of multiple sensors and articulators.
• To develop an architecture for artificial agents with a module that is responsible for multimodal articulation:
  • to implement action planning that is based on Finite State Automata that implement Sacks et al’s (1974) systematics of conversational turn-taking
  • to have the agent populations exhibit emergent turn-taking behaviour.

Our goals also include the evaluation of the elicited interactions as both artistic products
and scientific data:

• Artistically, to contribute to the study and implementation of interactivity.
• Scientifically, to evaluate our hypotheses regarding the attribution of agency and the human behaviours we anticipate to be elicited in interactions with low-fidelity agents, in the context of hetereogeneous agent populations.
• To develop multi-methodological evaluation that uses both qualitative and quantitative approaches.