A virtual human, augmented by a Facebook profile, is the subject of study in this paper. The paper is interesting on many levels, from the initial concept to the execution of the proof of concept. It also raises many questions.
The authors built a virtual human, endowed with Facebook information and powered by the Internet for both synchronous and asynchronous interactions. The paper posed two questions: (1) Could linking humans with social networking site information enhance human-virtual human interactions? (2) If so, how does it work?
About 20 subjects participated in two experimental interactive sessions in which they self-reported their observations and experiences. Results show that the subjects enjoy interacting with the virtual human called Sophie when she is informed by Facebook better than when she is not. They are also more likely to accept movie recommendations from her. Sophie herself not only generates speech, but also recognizes human emotions and reacts with body language, smiles, and hand waving. All of this supports what we might call a gut feeling about how a Facebook-fueled construct might appear--we just have not seen one until now.
This study is an important one, despite some flaws in the experimental design, the text-to-speech synthesizer, and the level of graphics (likely due to cost and time constraints). It is one of the first to put a virtual human face on top of social network information.