Anthropomorphism of computers: Is it mindful or mindless?

In analyzing the human tendency to treat computers as social actors (CASA), researchers tend to rule out the anthropomorphism explanation because anthropomorphism is understood to be "a sincere, conscious belief" that computers are human and/or deserving of human attributions. But, does anthropomorphism have to be necessarily mindful? Could it not also be a mindless tendency, especially given that most of us have somewhat long associations with our computers and have built human-like bonds with them? We examined these questions empirically by investigating whether the user tendency to treat computers as human beings is conscious (mindful) or non-conscious (mindless). We manipulated two variables (presence/absence of human-like agent and the low/high interactivity) on a health website and experimentally investigated whether they serve as anthropomorphic cues to trigger mindful attributions of human-ness to the website or mindless evaluations of the site in human terms. We found evidence for mindless anthropomorphism, with implications for user judgments of credibility of information on the site.