The Impact of Information and Communication Technologies on the Collective Intelligence of Human Groups
Final Report Abstract
A group’s collective intelligence describes its inherent ability to perform well on a wide variety of cognitive tasks. However, while collective intelligence in this sense was initially detected in the context of groups collaborating in a traditional face-to-face context, science today is indecisive whether the phenomenon emerges at all and specifically in contemporary digital contexts. We conducted a work program spanning 18 months, to analyze the emergence and structure of collective intelligence and also determine the impact of ICT-supported group work. In the following we summarize the key results of the research. We investigated how the existence and structure of collective intelligence in human groups is impacted by ICTs with varying degrees of media capacity. In our work program we analyzed the emergence and structure of CI in ICT-supported group work and found a single factor in Face-to-Face and Video treatments, but not in Audio and Text Chat treatments. - The ability of organizations to be skilled at team work and to seize opportunities to create high performance teams within the organization, is an important asset. Findings from our meta-analysis show, that the cumulative empirical evidence available to date refutes the prevailing theory of a single collective intelligence factor and suggests a more nuanced multi-factor structure. - There is a discrepancy in existing empirical research on CI: some research finds evidence of a general CI factor, while other research does not. Our research shows that groups exhibit CI when they work on well-structured tasks; groups that do well on some wellstructured tasks are likely to do equally well on other well-structured tasks. Ill-structured tasks require different cognitive processes from one task to the next, so groups are unlikely to exhibit CI when they work on ill-structured tasks: groups that do well on some illstructured tasks are no more likely to do equally well on other ill-structured tasks than some other randomly selected group. - We analyze the performance of hypothetical teams by ex-post aggregating the work of multiple individuals and comparing it to real teams. Thereby, we found convincing evidence that teams can work well without teamwork. - The utilization of collective intelligence via web-based applications is a promising approach for addressing wicked problems. The management of these information systems would benefit from evidence-based facilitation and improvement efforts, but to date, there is no suitable model to guide such efforts. We develop an assessment model for information systems that address wicked problems. The model was validated by assessing a web application dealing with the wicked problem of assessing the risks presented by global climate change. - We develop a decision-support tool that triggers and enriches facilitation actions in a crowd solving process in order to foster cross-fertilization as related field of CI. - We provide an overview of the literature on the wisdom of crowds and derive propositions. Further, we present the first general simulation model for aggregation of expert density judgments.
Publications
-
Clarifying the Structure of Collective Intelligence in Teams: A Meta-Analysis; Conference of Collective Intelligence, Carnegie Mellon University, June 13-14, 2019, Pittsburgh, USA
Graf-Drasch, V.; Gimpel, H.; Barlow, J. B.
