Many people have offered definitions of collective intelligence and related terms over the years (see definitions of collective intelligence at wikipedia). Rather than simply adopting one of these definitions here, however, we will first analyze the two words that make up the term:“intelligence” and “collective”.

What is Intelligence?

Even in fields that study intelligence, there are a variety of definitions of the term (see
definitions of intelligence from the psychology or artificial intelligence points of view).

Representative examples of the definitions include

• "Intelligence as the Tests Test It" (Boring, 1923).

• Turing test (Turing 1950)

• Intelligence is the “capability of a system to adapt its behavior to meet its goals in a range of environments” (Fogel, 1995, p. 22. See excerpt).

For now, we use a simple but very inclusive working definition: intelligence is "the ability to solve hard problems" (Minsky, 1985).

What is collective?

"Collective" has been defined as "of, relating to, characteristic of, or made by a number of people acting as a group" (American Heritage Dictionary ).

We suggest the following components in a working definition of "collective":

1 - One aspect of the dictionary definition--that of "acting as a group"--is what we have in mind.

2 - Since we are interested in both people and non-people entities (such as computers, animals), we generalize this definition to include non-people entites.

We suggest that the entities need not directly interact. For example, in a contest to guess the weight of a farm pig, the farmers' guesses could be independent of each other. But even there, there is a contest manager who takes on the role of aggregating the guesses so that all of them "act as a group."

What is collective intelligence?

Our current working definition of collective intelligence is:

Groups of individuals doing things collectively that seem intelligent.

Collective intelligence is a perspective, not just a phenomenon

Some phenomena in the world are clearly examples of collective intelligence: countries voting to elect a leader, companies deciding what products to make, and basketball teams deciding what to do with the ball. Even in these cases, observers may say that some groups acted “intelligently” and others didn’t. Other phenomena in the world (like water boiling or stars burning) are not easily analyzed as examples of collective intelligence at all.

In general, however, “collective intelligence” is a perspective that can be applied to many different kinds of phenomena. For instance, this perspective suggests another way of thinking about things like “organizational effectiveness”, “firm productivity”, “firm profitability”, “teamwork”, and “leadership.”

To apply the perspective of collective intelligence to a phenomenon, the following elements must be identified:

1. a group of actors (such as people, ants, or neurons)
2. a set of resources available to those actors (such as information and the capacity for taking certain actions)
3. a set of actions that the actors take (such as voters voting and the vote counting; or firms buying and selling, although in which case, there is only an "invisible hand" without an explicit central planner taking any action)
4. the collective results of the actions (such as winning a game, making money)
5. a way of evaluating the results (some outcomes are judged more “intelligent” than others).

Once these elements are identified, one can analyze how intelligently a given group acted given the resources it had, or one can imagine different ways of connecting the same actors and resources so that they could act more intelligently.

The perspective of collective intelligence suggests questions like the following:

• What would it mean for a group of people to be “intelligent”? For instance, if a single superhuman intelligence had access to all the knowledge and resources of a company like IBM or General Motors, what would it do? What strategies would it pursue? How quickly could it respond to changes in the marketplace? How productively could it use factories and money? How profitable would it be? And—most importantly—how closely could we approximate the behavior of this imaginary superhuman intelligence by cleverly connecting real people and computers?
• What can we learn from the ways human brains are organized that might suggest new ways to organize groups of people to perform intelligently? (And vice versa: What can we learn from the ways groups of people are organized that might help us understand how human brains are organized?)
• The field of artificial intelligence (AI) has, for decades, tried to create computer programs that can behave as intelligently as humans. From the traditional AI point of view, letting people help a program while it is running is considered cheating. But what if that were fine? What if the goal were to create combined human/machine systems that were more intelligent than either people or machines could be alone?

What is collective intelligence not?

Some people, when they hear the term “collective intelligence” assume that it implies individuals giving up their individuality to be somehow subsumed in a group. This is not what we mean.Collective intelligence, as we are defining and exploring it, is not about false consensus, cults, hive minds, or Groupthink. As described further in the section on "What factors inhibit collective intelligence," each of these phenomena actually represents a kind of collective stupidity or mediocrity. Similarly, although the Borg collective from Star Trek is an example of people and computers connected together to form a higher intelligence, it falls short of the true possibility of collective intelligence.

Note, that a simple answer to the first question above is that any formal group of people (e.g. modern organization) could sometimes be thought of as collective intelligence. But intelligence also implies rules that can be applied in many situations, and these rules form a certain identity. That's why the present IBM or GM can't be a good example of collective intelligence. The rules are the key point here.

As a statement of principle, we offer the following:

Collective Intelligence relies upon the individual knowledge, creativity, and identity of its constituent parts, and emerges from a synergy between them. In its highest forms, participating in collective intelligence can actually help people self-actualize while solving collective problems.