Presentation is loading. Please wait.

Presentation is loading. Please wait.

Do Agents and Avatars impact Group Processes? Do Agents and Avatars impact Group Processes? Lynsey Mahmood, Georgina Randsley de Moura & Tim Hopthrow University.

Similar presentations


Presentation on theme: "Do Agents and Avatars impact Group Processes? Do Agents and Avatars impact Group Processes? Lynsey Mahmood, Georgina Randsley de Moura & Tim Hopthrow University."— Presentation transcript:

1 Do Agents and Avatars impact Group Processes? Do Agents and Avatars impact Group Processes? Lynsey Mahmood, Georgina Randsley de Moura & Tim Hopthrow University of Kent KEY QUESTIONS? How do virtual group members impact group processes? Do virtual group members affect cooperative behaviour? Is there a difference in how individuals will respond to an Agent or an Avatar in a public good dilemma? THEORETICAL BACKGROUND 2.5 billion worldwide internet users as of 2012 (http://www.internetworldstats.com/stats.htm). 46% of U.S organizations asked use virtual groups (Society for Human Resource Management (SHRM), 2012). ‘Meeting’ online is now far easier- organizations can bring together the best people for the job, despite organisation and physical location (Lipnack & Stamps, 1999). Ingroup favouritism: people prefer members of their own group than outgroups, even in artificial groups (Webster & Wong, 2008). So it would be expected that investment to the central fund will be higher when the virtual group member is controlled by a human (ingroup) rather than a computer (outgroup). Human personalities can be applied to computers (Nass, Moon, Fogg, Reeves & Dryer, 1995) and individuals will mindlessly apply social rules to computers as they would humans (Nass & Moon, 2000) – suggesting there will be little difference in cooperation with the Agent or Avatar. Aim: to test these hypotheses, by comparing an Agent with an Avatar in an investment game, where the outcome would be affect the whole group. WHO TOOK PART? 189 undergraduate students (28 male, 160 female and one unspecified) Aged 18-38 years (M =19.15, SD =1.987) WHAT DID WE DO? Groups of six- sixth member was always virtual group member (pictured), either an Agent or an Avatar. Completed Horse trader problem (Zaleska, 1978) individually. Then heard virtual group members answer to the same problem. The virtual group member was pre-programmed to say: “I am an [agent/avatar]. I have seen the problem and I think the answer is [ten/twenty] pounds” Four conditions: Agent correct N=50, Agent incorrect N=54, Avatar correct N=45 and Avatar incorrect N=40. Four minute discussion leading to one group answer. Imagined public good dilemma scenario: Each person was endowed with a 300p personal investment. They were able to contribute any amount of this to a central fund, which at the end, would be doubled and shared out equally between group members (regardless of input). WHAT DID WE MEASURE? How much of the individual investment participants contributed to the group fund. Lynsey Mahmood lm455@kent.ac.uk @lynseymahmood What are Agents and Avatars?* Agent: Controlled by artificial intelligence which is not always correct but does learn Avatar: Controlled by a person who is not always correct but does learn *As described to participants A SNAPSHOT OF THE RESULTS A significant interaction was found for agent/avatar x individual matched scores (for horse trader problem) on own investment (F(1,95)= 4.196, p<.05) Avatar match M= 235.71 Avatar mismatch M=179.38 Agent match M=206.07 Agent mismatch M= 227.56 What does this mean? Participants invested more to the central fund when the virtual group member was an avatar and matched their answer to the horse trader problem. Participants invested the least when the virtual group member was an avatar that did not match their answer to the horse trader problem. What are the possible explanations for this? Similarity-attraction (Nass et al, 1995; 2000). Avatar controlled by human and matched answer so most similar to participant, leads to increased attraction, and willingness to work for group outcomes. Game theory (Artz and Gil, 2007). Trust agent if it will not gain from another's disadvantage, so trust that agent must be correct even if answer does not match own Decisions in computer mediated groups move further from individual to group decisions than in face-to-face groups (Siegel, Dubrovsky, Kiesler & McGuire, 1986). FUTURE DIRECTIONS Manipulate how much the virtual group member invests Compare participants expected virtual member investment with own Compare agent vs. avatar vs. human to test similarity-attraction Measure social identity and trust in virtual group members answerQUESTIONS? Why are we interested in one virtual group member? How does a virtual group differ from computer mediated communication in groups? Where could this research be applied practically?References –Artz, D. and Gil, Y. (2007). A survey of trust in computer science and the Semantic Web. Journal of Web Semantics, 5, 58-71. doi: 10.1016/j.websem.2007.03.002 –Internet world stats (2012) Retrieved from: http://www.internetworldstats.com/stats.htm on May 6th 2013 –Lipnack, J. and Stamps, J. (1999). Virtual Teams: The new way to work. Strategy & Leadership, 27, 14-18 –Minton-Eversole, T. (2012). Virtual Teams Used Most by Global Organizations, Survey Says. Retrieved from http://www.shrm.org/hrdisciplines/orgempdev/articles/Pages/VirtualTeamsUsedMostbyGlobalOrganizations,SurveySays.aspx –Nass, C. Moon, Y Fogg, B. J. Reeves and Dryer, D. C. (1995). Can computer personalities be human personalities? International Journal Human-Computer Studies, 43, 223-239. doi: 1071-5819/95/070223 –Nass, C. and Moon, Y. (2000). Machines and Mindlessness: Social Responses to Computers. Journal of Social Issues, 56, 81-103. –Siegel, J., Dubrovsky, V., Kiesler, S., & McGuire, T. W. (1986). Group processes in computer- mediated communication. Organizational Behavior and Human Decision Processes, 37, 157–187. –Webster, J. and Wong, W.K.P (2008). Comparing traditional and virtual group forms: identity, communication and trust in naturally occurring project teams. The International Journal of Human Resource Management, 19, 41-62. doi: 10.1080/09585190701763883 –Zaleska, M. (1978). Individual and group choices among solutions of a problem when solution verifiability is moderate or low. European Journal of Social Psychology, 8, 37-53


Download ppt "Do Agents and Avatars impact Group Processes? Do Agents and Avatars impact Group Processes? Lynsey Mahmood, Georgina Randsley de Moura & Tim Hopthrow University."

Similar presentations


Ads by Google