My agent as myself or another: effects on credibility and listening to advice

Journal article

Ian Li, Jodi Forlizzi, Anind K. Dey, Sara B. Kiesler
dppi, Designing Pleasurable Products and Interfaces, 2007 Jul 21, pp. 194-208

View PDF ACM DOI

Cite

APA Click to copy
Li, I., Forlizzi, J., Dey, A. K., & Kiesler, S. B. (2007). My agent as myself or another: effects on credibility and listening to advice. Designing Pleasurable Products and Interfaces, 194–208.

Chicago/Turabian Click to copy
Li, Ian, Jodi Forlizzi, Anind K. Dey, and Sara B. Kiesler. “My Agent as Myself or Another: Effects on Credibility and Listening to Advice.” Designing Pleasurable Products and Interfaces (July 21, 2007): 194–208.

MLA Click to copy
Li, Ian, et al. “My Agent as Myself or Another: Effects on Credibility and Listening to Advice.” Designing Pleasurable Products and Interfaces, July 2007, pp. 194–208.

BibTeX Click to copy

@article{ian2007a,
  title = {My agent as myself or another: effects on credibility and listening to advice},
  year = {2007},
  month = jul,
  day = {21},
  journal = {Designing Pleasurable Products and Interfaces},
  pages = {194-208},
  author = {Li, Ian and Forlizzi, Jodi and Dey, Anind K. and Kiesler, Sara B.},
  booktitle = {dppi},
  month_numeric = {7}
}

Abstract

People consider other people who resemble them to be more persuasive. Users may consider embodied conversational agents, or ECAs, to be more persuasive if the agents resemble them. In an experimental study, we found that users rated the persuasiveness of agents that resemble them higher than other agents. However, actual advice-taking diverged from this pattern; when users created the agents, users changed their choices less when interacting with the agents that resembled them. We conducted a follow-up study and found that resemblance and self-esteem affect interactions with agents that resemble users. We discuss the use of self-report and behavioral data in evaluations of agent interfaces and how agents that resemble users might foster particular social interactions with a system. We suggest that agents that resemble users may be more persuasive in advising users about their actions and decisions.