Collective intelligence and citizen experts: not democracy

Since the 1990s, a number of government agencies and other public authorities have experimented with increasing availability and possibilities for civic participation, using terms as e-democracy or e-government to categorise what they are doing.

Increased participation in this form has most commonly been seen as a way of increasing or improving input legitimacy in governance. In the example I will delve on here, output legitimacy is far more important.

The main idea behind collective knowledge systems can be summarised in the words of the French philosopher Pierre Levy: “Nobody knows all, everybody knows something, all knowledge resides in humanity.” Knowledge produced as an aggregate of what many people know is better than knowledge emanating from a single individual. Extrapolating, political decisions made on these premises are better than decisions made on information provided by a single bureaucrat or government expert.

However, no more than Wikipedia is the result of equal participation of peers, collective intelligence as a method for information gathering for government purposes will almost unavoidably gather expert knowledge effectively rather than democratise the production of knowledge.

The United States Patent and Trademark Office is since June 2007 using collective knowledge systems to evaluate patent applications in the Peer To Patent Project. The public is invited to scrutinise applications, and, according to the agency, the project will “empirically demonstrate the role that lay experts might play in improving decision making”.

The main problem of evaluating patent applications is to find out whether the patent brings something new to existing knowledge. The technical term is “locating prior art”. With the base of knowledge and the number of innovations growing exponentially, this becomes increasingly difficult, and it is impossible for a single bureaucrat to attain the knowledge needed in the process of establishing prior art in any given field. By opening up the process, someone or some people who have actually got that knowledge might be able to help.

Community reviewers register on the project site and have the possibility to go through applications and existing patents, scientific articles or other documents that might be said to define the prior art in the relevant field. This is a work carried out by a group of community reviewers and the participants discuss what document should be seen as important. In the end, what is agreed upon as the most important documents are sent to the Office.

For the sake of argument, I will not comment on the fast that the project seems to have failed: the number of community reviewers are at the time of writing alarmingly low. The relevant question is whether this is a form of democratised knowledge production.

A community reviewer must, in order to give relevant input to establishing the prior art concerning for instance computer software, have some form of pre-existing expert knowledge in the field. Potential participants might be anyone, but it doesn’t seem likely that anyone who hasn’t got the interest and knowledge would care the least for working - for free! - for a government agency. Also, the agency would hardly worry about any concerns raised by unknowing amateurs. The participating citizens must therefore be experts.

Apologists nevertheless maintain that the project is about enhanced democracy. I would rather hold that this is about expanding an elite of experts, not increasing political participation or contributing to equality. This does not mean that collective knowledge systems in public agencies shouldn’t be used. Only that they should not be used with the argument that they increase democracy.