ABSTRACT
This chapter discusses socially intelligent agents (SIA) in the context of human culture. The first part of the chapter provides an introduction to SIA research and focuses on the issue of realism versus believabilitynamely, to what extent SIAs need to imitate behavior and appearance of humans. The lifelike agents hypothesis is analyzed in particular from the perspective of social robots. I give examples showing that with respect to realism less might be more. Mori's "uncanny valley" analysis of believable robots is explained. It is shown that the issue of believability is particu larly crucial for technology whose purpose is to change the behavior, atti tudes, beliefs (or, more generally, minds) of humans, as it is the case for SIAs. Therefore, SIAs fall under the category of technology called persuasive technology. The second part of the chapter addresses primate cul tures, their phylogenetic origins, and how human culture is grounded in and shaped by cognitive capacities that the human species have evolved. I suggest that these cognitive capacities are important to consider in the design of SIAs. The chapter concludes by discussing the possible role of SIAs in the evolution of human cultures. I argue that SIAs can play a role in preserving cultural diversity-a huge technological challenge, but a possibil ity to make SIA technology more humane by acknowledging the individual and cultural roots of people, rather than treating them as an increasingly homogeneous group of users.
