The crowd can be smart and the crowd can be stupid. Part of the discussion on Brittanica deals with what the advantage is of social software and iCrowds. And when results are stupid or irrelevant. IS there for example a thing like a citizen scientist?
Open source seems to work great: Linux, Samba, Apache are all examples where the software has reached tremendous stability because it is open. Everybody in the world can test it, can comment on features and can look at the code to try to make it better. I think most will agree that in these cases no group of “professional specialists” would have done any better.
On the other hand we have Wikipedia that sort of combines the best and worst of information. And it can be hard to distinguish between good and bad. Parroting each other through blogs and fora’s creates information that looks like the truth because it is amplified a thousand times (you can Google it and a hundred links say the same “truth” but in reality it may be just one person who is the source and who might be wrong. A thousand consistent lies remains a lie. One of the reasons we think information that is found many-fold is true is because in the old days much of this information was from different sources. Due to the Internet and the free copy and distribution, information is replicated with the speed of lightning. This speed of replication leads to a loss in diversity of sources since it is much more convenient to copy based on what Google delivers. This process is of course self propelling.
One of the reasons for this difference is the ease how we can distinguish right or wrong. When using the Apache server I know when it crashes or is vulnerable for attacks. And I know when it is fixed (when it does not crash or does like advertised). also, each person looking at the application or code does this on it’s own. With information in Wikipedia this is not so easy to determine without thorough research based on “peer reviewed” sources.
As for the citizen scientist: there may certainly be a place in science for the pro-am model. A much mentioned example is the use of amateur astronomists that see a lot a things that the professionals missed because you can’t look at the whole sky. But collection of data is only a part of science. Building theories and designing experiments is something altogether different. On a photograph it is easy to see if it was a supernova. Building and proving a theorie is much less clear and needs the input and critical review of many experts.
Combining diversity is the key to iCrowds, not replication of the same.