COLLECTIVE INTELLIGENCE PROCESSES AND THEIR
INFLUENCE ON THE DYNAMICS OF INFORMATION
DIFFUSION ON THE WEB
Leonardo Caporarello and Luca Ongaro
Learning Lab, SDA Bocconi, Bocconi University, Via Bocconi 8, 20136 Milan, Italy
Keywords: Collective intelligence, Self-organization, Internet, Social web, Information network, Information overload.
Abstract: The so-called Social Web produced a paradigm shift toward a distributed model of information production
and diffusion, as well as an exponential growth of the total amount of information circulating in digital
format. The environment in which this information diffuses is a distributed global network of individuals
and organizations acting at the same time as information producers and consumers. The emergent dynamics
of self-organisation of this network fall into the realm of collective intelligence, and their study is essential
in order to understand their influence on the pattern of information diffusion and cultural development.
1 INTRODUCTION
The evolution of the Internet toward the so-called
Social Web (Gruber, 2008) produced several well-
known important implications on human cultural
development and on access to information (O’Reilly,
2005). The number of weak social ties and
information sources that an individual can manage
dramatically increased, as connections are no more
geographically bounded. Moreover, the easiness of
information transfer on the Web is continuously
rising up. Also, new technologies and pattern of use
of the Web increasingly allowed for a gradual
change of paradigm toward a distributed production
of information, with no formal distinction between
producers and consumers of information.
Thus, the circulation of ideas at a global level,
and the possibility to contribute to the creation and
management of knowledge in a distributed fashion is
determining the astonishing rate of Web content
production. The global amount of information
produced and stored in digital formats encountered a
tenfold increase in the last ten years, with a +60%
year-to-year growth rate (Gantz et al., 2007). In
absence of a mechanism of information filtering, we
could expect the problem of selecting the relevant
information, and discriminating it from “noise”, to
become increasingly complex. This problem is often
referred to with the expression “information
overload”. Some authors propose that the exposure
to a large amount of diverse information on the
Web, with varying degree of relevance, do produce
some measurable negative effects on the ability to
concentrate on tasks and on the exercise of deep
cognitive thought in the consumption of information
(Carr, 2008).
Some research evidences show that Internet users
are growing at a rate of 100 millions new users per
year (ITU Database, 2010), and that people are
spending more and more time on-line. This leads us
to think that the perceived benefit associated to the
social Web outweighs the cost introduced by
problems related to information overload. Due to the
decentralized nature of the Web, any mechanism
that mitigates the issue of information overload is
expected to come from a distributed effort.
According to the above considerations, our
research question is whether emergent and non-
supervised distributed mechanisms, falling within
the realm of collective intelligence (Bonabeau,
2009), effectively influence the dynamics of
information diffusion on the Web, balancing out
information overload. Moreover, if such a
mechanism is found, we intend to identify its main
implications on cultural development at both macro
and micro levels.
501
Caporarello L. and Ongaro L..
COLLECTIVE INTELLIGENCE PROCESSES AND THEIR INFLUENCE ON THE DYNAMICS OF INFORMATION DIFFUSION ON THE WEB.
DOI: 10.5220/0003445905010504
In Proceedings of the 1st International Conference on Cloud Computing and Services Science (CLOSER-2011), pages 501-504
ISBN: 978-989-8425-52-2
Copyright
c
 2011 SCITEPRESS (Science and Technology Publications, Lda.)

2 THEORETICAL
BACKGROUND
In order to investigate the research question, a
theoretical background on which to base our analysis
is needed. In particular, it is necessary to provide:
• A formal definition of self-organizing
processes, emergent phenomena and collective
intelligence
• A model explaining the nature of information
and the dynamics of distributed information
production and cultural development
• A model capable of effectively describe the
dynamics of information diffusion on the
Internet
A brief discussion introducing such definition
and models is thus required to elaborate on the
research question.
2.1 A Definition of Collective
Intelligence
Collective intelligence is a cognitive process
emerging at a systemic level from the interaction at
the micro-level between many distributed agents
(Levy, 1997). It represents a subset of the broader
concepts of emergent phenomena and mass
behavior, both involving surprising ordered patterns
produced by distributed agent-based systems,
generally non predictable from the study of a single
agent. Collective intelligence is thus an emergent
phenomenon involving some kind of decision-
making happening at the system level.
There are numerous examples of collective
intelligence in nature, such as the foraging behavior
of ant colonies or the complex schooling behavior of
fish without an individual leader. Collective
intelligence processes are common in human
societies too, an example of primary importance
being the market, where the price and allocation of
goods and assets is not the outcome of a central
decision, but rather of the combined judgements of
many independent individuals.
2.2 Memetics: A Model of Distributed
Cultural Development
Memetics is a theory aimed at explaining how
distributed and generally non-supervised social
systems can produce and develop ideas, culture and
also more trivial phenomena like fashion and trends.
It was first proposed by Richard Dawkins in the final
chapters of the book The Selfish Gene (1976), and
subsequently developed by other authors (i.e.
Dennet, 1992). The basic idea is to apply the theory
of evolution to culture, postulating the existence of
stable units of information, called memes, which
play a role analogue to the one of genes in biological
evolution.
In order to trigger an evolutionary process, an
environment needs to provide at least a mechanism
of replication capable of producing copies of
entities, a mechanism of variety generation
producing entities with novel characteristics, and a
process of selection increasing the chances of
replication of entities that better satisfy a “fitness”
requirement. Memetics identify the entities being
replicated with memes and the replication
mechanism with human communication: each time a
transfer of information happens between two people,
by means of explicit communication or by
observation of actions or behavior, ideas and their
building blocks (the memes) are replicated, so that at
the end of the process new copies of the original
ideas are created and hosted in the brain of the
information receiver. Imperfections of
communication, subjective interpretation of ideas
and recombination with other concepts all create
variation in the memetic pool. Finally, a selection
mechanism is introduced by the fact that some ideas
are more likely to be spread than others. The
“fitness” criteria of memes is not necessarily
bounded to advantages for the humans that carry
them, but still ideas that provide a benefit for
societies that adopt them are more likely to be
preserved and spread, while those that are harmful
tend to be discarded in the long run.
Memetics makes an attempt to explain how
complex cultural systems can arise from the
interaction of large communities of humans, and the
theory can be easily applied to cultural development
in the Internet era. The highly codified nature of
information on the Web and the fact that it can be
efficiently copied, stored and transferred make the
metaphors introduced by Memetics even more
applicable.
3 THE INFORMATION
NETWORK: A MODEL OF
INFORMATION DIFFUSION
ON THE WEB
The Web is essentially a networked structure of
resources and hyperlinks between them, and as such
CLOSER 2011 - International Conference on Cloud Computing and Services Science
502

can be easily represented as a directed graph, the
pages being the vertices and the links being the
directed edges. The evolution of the Internet as a
more and more social phenomena, with sites and
content often representing the direct emanation of
individuals or organizations being at the same time
producers and consumers of information, enables us
to apply the directed graph metaphor at a higher
level of analysis. The social Web, in fact, makes
information production and publication widely
accessible, and widespread technologies like
bookmarking of sites, RSS feeds, social networking
applications and blogging platforms make it easy to
keep track of interesting sources of information.
Thus, it is possible to model the social Web as a
directed graph in which the vertices or nodes are the
people and organizations producing and consuming
information (to which we will generally refer as
agents), while the arcs or directed links express the
relationship “A follows B” or “B is a source of
information for A”, A and B being two nodes.
The act of starting to follow a certain source of
information is generally a deliberate decision, and
from the moment in which that link is created there
is a directed flow of information happening between
those two nodes. Links in this model express also a
flow of information, and possibly an influence
exercised by one node on another, but generally not
a “friendship” relationship, nor any other kind of
mutual personal relationship between the two nodes.
For these reasons this model, which we call here
“information network”, should not be confused with
a social network.
The information flowing through the network
can be thought as composed of conceptual units, or
memes, and the information network is ultimately
the environment in which the kind of evolutionary
cultural processes described by Memetics take place.
Nodes are in fact acting as repositories of memes,
and units of information are replicated and
transferred through the arcs of the information
network. Information that is more “fit” will have
more chances to be propagated from link to link,
with nodes acting as information relays, while
information with low value for nodes in a certain
region of the network will tend to be blocked. In this
process, information is often processed, and the
memes of which it is composed are recombined and
selected.
It is especially interesting to analyse the
dynamics of evolution of an information network.
There is in fact an upper bound to the number of
nodes that a single node can follow, since following
too many sources would cause a node to be unable
to make use of all the information it receives. Thus,
any node will try to maximise the value of the
information it receives by choosing to follow those
nodes that produce or diffuse information that is
particularly interesting and valuable from its
particular perspective. The value of information is
not absolute, but different from node to node, as
different people have different capability to make
use of specific information, as well as different taste.
Furthermore, agents acting as nodes in the network
have bounded rationality, and most importantly a
“partial horizon”, since they cannot observe the
whole information network, and thus they cannot
choose a globally optimal solution to their problem
of selecting the right sources/nodes. Instead, they
progressively change their set of sources with steps
of improvement, following new nodes that provide
information of high value when they get to know
them from their current sources, and ceasing to
follow the least interesting ones. As a result of the
distributed effort of each node trying to optimize its
set of sources, the whole information network
evolves its structure, shaping the paths through
which different kind of information diffuses. If the
structure of the information network is found to
evolve toward an ordered and non-random
equilibrium, that same equilibrium is by definition
the outcome of a collective intelligence process,
since - as said - no single node has the authority or
the possibility to globally shape the network, and the
global structure is the product of bounded decision
taken by nodes at a micro level.
4 RESEARCH DIRECTIONS
The authors propose a quest for emergent collective
intelligence processes on the Web that may shape
the way in which information diffuses and culture
evolves. The research will involve an analysis of
existing “natural experiments”, on which to make
quantitative measurements and test the proposed
model of information network. One promising
research methodology would be to make use of web
crawlers to collect data on relevant indicators of
users’ behavior in the information network, as well
as measures of the efficiency of the information
routing. In other words, a first aim will be to
measure how good Internet communities are at
collectively self-organize to improve the signal-to-
noise ratio and to deliver information quickly to the
users who can benefit most from it, while filtering
out non-relevant or flawed information.
COLLECTIVE INTELLIGENCE PROCESSES AND THEIR INFLUENCE ON THE DYNAMICS OF INFORMATION
DIFFUSION ON THE WEB
503

A second research direction will be the analysis
of the dynamics of evolution of information
networks. For this purpose it will be particularly
important to build realistic agent-based simulation
models, in order to test the impact of various factors
on the pattern of self-organization of the whole
network.
Hopefully, the analysis of emergent collective
intelligence processes in information networks will
provide practical implications and applicable
considerations on how to positively influence their
evolution toward an optimal configuration. The
knowledge resulting from this kind of research may
thus contribute to the field of distributed information
systems design.
REFERENCES
Bloom, H., 2001. Global Brain: The Evolution of Mass
Mind from the Big Bang to the 21st Century, Wiley
Books
Bonabeau, E., 2009. Decisions 2.0: The Power of
Collective Intelligence, Sloan Management Review,
50(2)
Border, A., 2006. Collective intelligence: a keystone in
knowledge management, Journal of knowledge
Management, 10(1)
Carr, N. G., 2008. Is Google Making Us Stupid?, The
Atlantic, Jul/Aug 2008 issue
Dawkins, R., 1976. The Selfish Gene, Oxford University
Press (Book)
Dennet, D. C., 1992. Consciousness Explained, Back Bay
Books
Gantz et al., 2007. The Diverse and Exploding Digital
Universe: An Updated Forecast of Worldwide
Information Growth Through 2011, ITC White Paper,
http://www.emc.com/collateral/analyst-reports/expan
ding-digital-idc-white-paper.pdf
Gruber, T., 2008. Collective knowledge systems: Where
the Social Web meets the Semantic Web, Web
Semantics: Science, Services and Agents, 4(13)
Levy, P., 1997. Collective Intelligence: Mankind's
Emerging World in Cyberspace, Perseus Books,
Cambridge MA
O’Reilly, T., 2005. What is Web 2.0: Design Patterns and
Business Models for the Next Generation of Software,
http://www.oreillynet.com/pub/a/oreilly/tim/news/200
5/09/30/what-is-web-20.html
CLOSER 2011 - International Conference on Cloud Computing and Services Science
504