Collective intelligence

Strategic Finance, Nov, 2006 by Michael Castelluccio

HOW MANY TIMES have you been in this situation? The chairs are all in a circle, and the moderator smiles as he extends both hands, palms up, inviting all to enter a groupspeak for the next two or three hours. When you were back in school, everyone sort of understood that this kind of exercise would not likely get much beyond pooling your ignorance, but does it work any better now that you and your peers have more experience and are more chatty? Is there a "collective intelligence" that can be created in groups--an intelligence that far outstrips any of the individuals in the group?

A group of academics at MIT seem to think that it's possible to get more than just the sum of the individuals' abilities when several put their heads together to solve a problem. The MIT group, called the Center for Collective Intelligence, represents a wide spectrum from the various schools at the Institute, including the Sloan School of Management, the Computer Science and Artificial Intelligence Laboratory (CSAIL), the famous Media Laboratory, McGovern Institute for Brain Research, and the MIT Leadership Center. On a recent Friday the 13th in October, the group met to announce a year-long project to investigate collective intelligence (CI). The tangible goal is to collectively write a book titled We Are Smarter than Me, to be published by Pearson in 2007. There will also be a conference called Community 2.0 next spring.

Can We Think as a Group?

As the various speakers presented their individual thoughts at the launch, several definitions and approaches emerged.

Claude Canizares, VP for Research at MIT, outlined the challenge. Basically, it's how can we take huge amounts of information and make sense of it? He listed all the departments involved in the project, explaining the advantage that MIT has in its long tradition of having porous boundaries between disciplines. In this case, those working on the project range from neuroscientists to business professors. The benefits of collective intelligence, he explained, could be significant. "It's not inconceivable that the kind of work that would come out of CCI (Center for Collective Intelligence) would really cause a revolution in the way people work and think."

Thomas Malone, director of CCI, provided a working definition of collective intelligence: "groups of individuals doing something collectively that seems intelligent." Seen this way, CI has been around a long time because families, countries, even beehives and ant colonies, do things together that seem intelligent, like form governments or seek food. But in the last few years, he explained, we have seen some new examples of group intelligence tied to new technologies. Google, for instance, takes the collective knowledge of millions of online people and organizes it into directed answers. Wikipedia is the product of thousands of editors. Alph Bingham's company, InnoCentive, Inc., uses a network of scientists to work on problems that are presented to the community. Bounties are offered for solutions, and the network is free to try and win them. Open Source software is developed by communities, not individuals, and Malone said he thinks all of these are just the beginning of a much larger collective movement. The Internet makes it possible to "harness huge numbers of people on a scale never before possible and connected in very different ways than has ever been possible before." To take advantage of the potential, we need greater understanding of what can be done, he explained, and that is the reason for the CI project. The key question then will be, "How can people and computers be connected so that collectively they act more intelligently than any individual, group, or computer has ever done before?"

To proceed, Malone explained they will need three types of experiments: case studies; they will have to create new examples of CI; and there will have to be systematic studies of the examples. And, finally, new theories are needed to tie everything together.

Malone finished his remarks with descriptions of two extremes, neither of which he believes is credible. The first is those who think that CI is kind of magical, and all you need to do is gather a group and results will emerge. At the other end are those who think that nothing is possible without a strong central organizing direction (intelligence) imposed on the groups.

In his presentation, Alph Bingham asked the question, "How does creativity fit in to CI?" He then described two classic approaches to scientific problems. Thomas Edison, on his way to inventing the light bulb, tried and tested numerous elements, containers, and gases until he produced his incandescent bulb. His method was the classical hypothesize, experiment, evaluate, and re-hypothesize. The Greek Archimedes had the problem of measuring the volume of an irregular crown in order to get its density. He wanted to see if the gold was pure or mixed with other metals--rather, it was the Tyrant of Syracuse who wanted to know if he had been ripped off. But the scientist's solution did not result in a mathematical formula for irregular shapes. While taking a bath, Archimedes noticed that he displaced a measurable amount of water, and soon he was running down the street with the solution and his famous "Eureka!"