Information markets, prediction markets, swarm intelligence, the wisdom of crowds – even in 2009, mentioning collective intelligence to decision-makers can result in raised eyebrows and inquisitive glances. It also begs a question: what actual use is it for industry today? To look at the issue of collective intelligence, people from science and academia joined managers and technology users at a top-level research workshop in March. The SMI Spring Workshop took place in Stuttgart and was organized by Prof. Dr. habil. Andreas Aulinger and Max Pfeiffer on behalf of the Steinbeis School of Management and Innovation (SMI) at Steinbeis University Berlin (SHB) and the Ferdinand Steinbeis Institute (FSTI).
Both scientists are closely involved in projects looking at the promotion of collective intelligence. Some of the issues they have stumbled across were discussed with the experts from science and industry at the workshop. One starting point of collective intelligence research was the work of British scholar Francis Galton. On a visit to a cattle market one day (the “West England Fat Stock and Poultry Exhibition”), Galton monitored a competition to guess the weight of a huge ox. For a sixpence, visitors could submit their guess on a scrap of paper. Around 800 people tried their luck. After evaluating the guesses, Galton discovered that the average of all guesses was less than 0.1% from the actual weight of the ox: its true weight was 1198 lb, the average of all guesses was 1197 lb.
In another area of collective intelligence, researchers pinpointed the amazing ability of swarms of animals. Despite the lack of hierarchies and the limited cognitive ability of certain types of animals (such as fish, ants and birds), entire swarms react in an extremely intelligent manner based on a small selection of instinctively adhered to behavioural patterns – thus the common term swarm intelligence is used. It's often equalized to collective intelligence.
In the literature, however, a clear distinction is made between the two phenomena. For example, Aulinger refers to the ox story as an example of collective intelligence involving “many unaffected”. With swarm intelligence he describes it as collective intelligence of “many affected”. In the former example collective intelligence is the result of many agents providing predictions or descriptions of a situation independently of one another, culminating in an average value. In the latter, collective intelligence results from agents reacting directly to each other and a small number of instinctively followed behavioural patterns. The SMI spring workshop examined both forms of collective intelligence and went on to ask what could be applied to modern companies and what fundamental attributes of collective intelligence exist.
In total, 14 experts came from all over Germany to give a lecture at the SMI spring workshop and discuss their research and practical interests in collective intelligence.
Professor Ulrich Krause from Bremen University proposed a field experiment on swarm intelligence and made a link to mathematical computations based on averaging. Stephan Stathel from the IT research center at the University of Karlsruhe looked at the value of innovation scoring in information markets. Max Pfeiffer, Director of the Ferdinand Steinbeis Institute and doctoral student at Steinbeis University, explained the significance of information markets as one part of prediction markets. Information markets try to make knowledge of the individual accessible by using the incentive structures of the community – e.g. departments, companies or public bodies – and, by making predictions, package this into aggregated opinion. These predictions can then be used within companies as a benchmark for new ideas and to support strategic decisions. Christian Slamka from Frankfurt University presented idea markets as another area of use. Laura Miller, a Steinbeis University doctoral student from Rofin Sinar Laser GmbH showcased her experience with processes that use collective intelligence to make market predictions for SMEs.
It is not only companies that use the inputs of individuals to form intelligent predictions. Dr. Oliver Märker from Zebralog GmbH & Co. KG outlined how the City of Cologne’s “citizen’s budget” initiative uses new media in the form of e-participation to improve public participation and use collective intelligence to optimize the allocation of resources.
To add everyday examples of applications to the workshop, three providers working in information markets were invited to talk about their experience in this area. Dr. Bernd Ankenbrand, managing director of Witten/ Herdecke University’s knowledgelab and Gexid GmbH reported that even the US scientific community – a driving force and hotbed of instruments promoting collective intelligence – still has no clearly established model. There may be dozens of live examples from companies like Google, Hewlett Packard and Microsoft, that show that information markets can possess a wealth of predictive intelligence, but an analytical model has yet to be produced.
Drawing on examples with the Analyx platform, Georg Peller and Caroline Rudzinski from Management Zentrum Witten also showed that retailers are already successfully using predictive markets. However, despite proven success, generally applicable success factors have yet to be identified. As well as looking at the use of information markets, Dr. Christoph Hartl described how the German army is also trying to speed up the process of supplying clear situation reports during complex missions by linking up technologies and infrastructures more closely. Andreas Schäfer, from Fraunhofer IAIS, drew on a number of examples to demonstrate that consumers are already unknowingly part of collective intelligence thanks to data mining. Anonymised positioning data generated by mobile devices with navigation systems is merged to provide traffic updates in real time. Collective intelligence methods are especially useful when making predictions in real time. Possible cost benefits – and a strong ability to stimulate those involved – spell much potential for this approach to information gathering.
As part of the SMI Spring Workshop, Swarmworks Ltd. provided some innovative ways to evaluate the results of the workshop, connecting large groups interactively and intelligently via connected workstations. A brainstorming session resulted in seven core fields where collective intelligence is expected to be applied:
In-house applications (fostering motivation, change management, recruitment)Idea management/screeningSocial applications (budgeting, natural resource management)Gathering feedbackDecision making (forming opinions in real time, decision preparation)General information management (identifying experts, connecting knowledge management systems)
The complete findings of the SMI Spring Workshop on collective intelligence will be released in Autumn 2009 in the conference minutes published in the Steinbeis Edition. Edited by Prof. Dr. habil. Andreas Aulinger and Max Pfeiffer, the volume provides an excellent summary of the current status of collective intelligence.