Computers Are Social Actors: A New Paradigm and Some Surprising Results

Cliff Nass, Stanford Dept. of Communcation

Seminar on People, Computers, and Design
Stanford University November 4, 1992


In this talk, I'll present the results from a number of experimental studies that demonstrate that individuals can be induced to use a number of social rules in evaluating the performance of computers and other technologies; that is, computer-sophisticated individuals exhibit anthropomorphism. We were able to demonstrate via laboratory experiments that when subjects evaluate computers, they use rules such as "praise of others is more valid than praise of self" and "praise of others is friendlier than is praise of self, while derogation of others is friendlier than is derogation of self." We also demonstrate that people use the ideas of role-specialization to interpret the behavior of technologies, perceiving the content portrayed on televisions to be better and more homogeneous on televisions that are used to portray only one type of content than on televisions that portray multiple types of content. Implications for interface design are discussed, including the use of voice to create low-overhead agents and the dangers of general-purpose software.


Clifford Nass is presently an assistant professor at Stanford University in the Department of Communication, with courtesy appointments in Sociology and Values, Technology, Science, and Society. He received his B.A. cum laude in Mathematics and his Ph.D. in Sociology, both from Princeton University. Nass worked for a number of years as a computer scientist at the IBM Research Center and Intel corporation, designing graphics and database algorithms. His primary areas of interest include social-psychological models of the human computer interaction ("Computer as Social Actor"), technology (particularly computer technology) and society, and statistical methodology. His publications have appeared in such journals as Communication Research, Human Communication Research, Administrative Science Quarterly, and IBM Technical Disclosure Bulletin.


Titles and abstracts for all years are available by year and by speaker.

For more information about HCI at Stanford see

Overview Degrees Courses Research Faculty FAQ