The social dynamics of human-robot interactions: Generalizations from real people to a comprehensive sample of social robots

Transcription

1 The social dynamics of human-robot interactions: Byron Reeves Jeff Hancock Sunny Liu Generalizations from real people to a comprehensive sample of social robots

2 Social Robots Socially evocative Socially situated Sociable Socially intelligent more variance exists between different examples within a category of social technology than between human responses to any one example in the category Problem of Stimulus Sampling

3 Study 1 - All (most) the Social Robots 1) learn about the social dynamics of human perceptions of robots 2) but do so considering human perceptions across as large a range of robot examples as possible.

4 Study 1 - All (most) the Social Robots Inclusion Criteria (1)a social robot must be named in the article as the focus of the research (2) there must be a picture(s) of the robot in the article or in supplemental materials (3)the relationship of the social robot with people must be discussed and studied.

5 Study 1 - All (most) the Social Robots Total Articles on Google Scholar (light green), Satisfying Inclusion Criteria (dark green) TOTAL ARTICLES: 6960 TOTAL INCLUDED: 1471

6 N = 342

7 Study 1 - All (most) the Social Robots Stanford Social Robot Collection - images of each robot - database of all related papers

8

9

10 Stereotype Content Model

11 Stereotype Content Model FROM: FISKE, CUDDY, GLICK, AND XU, 2002

12 Stereotype Content Model Active Facilitation Admiration Pity Passive Harm Passive Affiliation Contempt Envy Active Harm FROM: FISKE, CUDDY, GLICK, AND XU, 2002

13 Stereotype Content Model Active Facilitation Admiration Pity Passive Harm Passive Affiliation Contempt Envy Active Harm FROM: FISKE, CUDDY, GLICK, AND XU, 2002

14 Study 2 - Warmth and Competence Dimensions (not at all) (extremely) N = 3920 Turkers Each robot was rated by 10 turkeys competent tolerant confident warm independent good natured competitive sincere intelligent

15 Study 2 - Warmth and Competence Dimensions Item Component 1 Warmth Component 2 Competence Mean (SD) competent (.38) tolerant (.37) confident (.38) warm (.55) independent (.36) good natured (.47) competitive (.40) sincere (.40) intelligent (.41) 35.8% 32.0% variance accounted for

16 Study 3 - Features Predicting Warmth and Competence

17

18 Features Head eye:head ratio head:body ratio has face has vision Skin Type & Shape plastic metal fur silicone mechanics visible animal shape feet Communication Ability has speech has digital presentation Motion degrees of freedom locomotion bipedal Gender gender displayed masculinity femininity Age developmental category age

19 Study 3 - Features Predicting Warmth and Competence Features Warmth Competence Head eye:head ratio 0.14* 0.03 head:body ratio has face has vision 0.23** 0.08 Skin Type & Shape plastic metal fur ** silicone mechanics visible -0.17** animal shape feet Communication Ability has speech has digital presentation Motion degrees of freedom *** locomotion bipedal Gender gender displayed masculinity femininity Age developmental category age -0.16** 0.08 Adjusted R

20 Study 3 - Features Predicting Warmth and Competence Features Warmth Competence Head eye:head ratio 0.14* 0.03 head:body ratio has face has vision 0.23** 0.08 Skin Type & Shape plastic metal fur ** silicone mechanics visible -0.17** animal shape feet Communication Ability has speech has digital presentation Motion degrees of freedom *** locomotion bipedal Gender gender displayed masculinity femininity Age developmental category age -0.16** 0.08 Adjusted R

21 Study 3 - Features Predicting Warmth and Competence Features Warmth Competence Head eye:head ratio 0.14* 0.03 head:body ratio has face has vision 0.23** 0.08 Skin Type & Shape plastic metal fur ** silicone mechanics visible -0.17** animal shape feet Communication Ability has speech has digital presentation Motion degrees of freedom *** locomotion bipedal Gender gender displayed masculinity femininity Age developmental category age -0.16** 0.08 Adjusted R

22 Study 3 - Features Predicting Warmth and Competence Features Warmth Competence Head eye:head ratio 0.14* 0.03 head:body ratio has face has vision 0.23** 0.08 Skin Type & Shape plastic metal fur ** silicone mechanics visible -0.17** animal shape feet Communication Ability has speech has digital presentation Motion degrees of freedom *** locomotion bipedal Gender gender displayed masculinity femininity Age developmental category age -0.16** 0.08 Adjusted R

23 Stereotype Content Model

24 Stereotype Content Model Active Facilitation Admiration Pity Passive Harm Passive Affiliation Contempt Envy Active Harm FROM: FISKE, CUDDY, GLICK, AND XU, 2002

25

26 from: Todorov (2008)

27 The social dynamics of human-robot interactions: Byron Reeves Jeff Hancock Sunny Liu Generalizations from real people to a comprehensive sample of social robots