230
11 The Self-Organizing Society
manage our future in an increasingly complex society, it is important to encourage
and consider multiple perspectives. A symbiotic relationship with digitally literate
citizens, customers and users is key to success. Our society can only live up to its
capacity, if it makes best use of the skills, ideas and resources of its citizens. It
will be of strategic importance to offer participatory social, economic, and political
opportunities. In future, those societies will be leading, which manage to create a
win-win-win situation between businesses, citizens, and state.
Generally, to reap the benefits of the digital revolution, I recommend engaging
more in distributed storage, processing, and control. In complex systems, a decentralized kind of organization can be superior to a centralized system.
18 This is surprising,
but a centralized approach often ignores local knowledge (since it is usually not
possible to centrally process all local information). Bottlenecks such as insufficient
processing power and data transmission rates are limiting factors, also in future. The
use of local knowledge, in contrast, allows decentralized approaches to thrive.
The local interactions between the many components of a complex dynamical
system can produce emergent structures, properties, or functionalities based on selforganization. However, as traffic jams, crowd disasters, financial crises and “tragedies
of the commons” show, self-organization does not always create desirable outcomes.
Nevertheless, these phenomena are now reasonably well understood and can be replicated using mathematical models and computer simulations. Those simulations tell
us that the negative outcomes of self-organization can often be avoided by changing
the interaction rules, i.e. the mechanisms by which the components of the system
interact. In some cases such as traffic flows or the financial system, simply altering
the system’s parameters (such as the vehicle density or interest rate) can avoid or
reduce undesirable consequences. In fact, while the “invisible hand” (which may be
seen as another term for “self-organization”) often fails if network effects or externalities matter,
19 we can now overcome such failure. 300 years after the concept of
the “invisible hand” was invented, we can let it work for us! The sensor networks
behind the “Internet of Things” enables us to realize Adam Smith’s brilliant vision of
self-organizing systems for the first time in human history. While this is an unprecedented opportunity to make our increasingly complex world manageable again, we
need to fundamentally change the way we think about global governance.
18 Couldn’t we emulate a decentralized system by a centralized one? This seems plausible, but
besides being more expensive, it is not obvious which data is relevant and what data obfuscates the
truth. The relevant signal in one application might be noise in another application and vice versa.
Therefore, in contrast to the conventional wisdom on Big Data, less information can be better. Let
me give an example. As I previously discussed, it is possible to understand the spread of epidemics
combining small datasets of infection cases and airplane trips. This approach performs better than
a large centralized system powered by Big Data such as Google Flu Trends. The reason is that too
much data produces problems such as “spurious correlations” and “over-fitting”, i.e. the fitting of
random, meaningless patterns. In other words, results of Big Data analytics may be irrelevant and
misleading, which can cause bad decisions.
19 Note that eliminating network effects by world markets and free trade is not a perfect solution, as
it eliminates the niches which are needed for innovation. Moreover, externalities will always play
a role when production takes place in densely populated areas or in vulnerable ecosystems.
11 The Self-Organizing Society
manage our future in an increasingly complex society, it is important to encourage
and consider multiple perspectives. A symbiotic relationship with digitally literate
citizens, customers and users is key to success. Our society can only live up to its
capacity, if it makes best use of the skills, ideas and resources of its citizens. It
will be of strategic importance to offer participatory social, economic, and political
opportunities. In future, those societies will be leading, which manage to create a
win-win-win situation between businesses, citizens, and state.
Generally, to reap the benefits of the digital revolution, I recommend engaging
more in distributed storage, processing, and control. In complex systems, a decentralized kind of organization can be superior to a centralized system.
18 This is surprising,
but a centralized approach often ignores local knowledge (since it is usually not
possible to centrally process all local information). Bottlenecks such as insufficient
processing power and data transmission rates are limiting factors, also in future. The
use of local knowledge, in contrast, allows decentralized approaches to thrive.
The local interactions between the many components of a complex dynamical
system can produce emergent structures, properties, or functionalities based on selforganization. However, as traffic jams, crowd disasters, financial crises and “tragedies
of the commons” show, self-organization does not always create desirable outcomes.
Nevertheless, these phenomena are now reasonably well understood and can be replicated using mathematical models and computer simulations. Those simulations tell
us that the negative outcomes of self-organization can often be avoided by changing
the interaction rules, i.e. the mechanisms by which the components of the system
interact. In some cases such as traffic flows or the financial system, simply altering
the system’s parameters (such as the vehicle density or interest rate) can avoid or
reduce undesirable consequences. In fact, while the “invisible hand” (which may be
seen as another term for “self-organization”) often fails if network effects or externalities matter,
19 we can now overcome such failure. 300 years after the concept of
the “invisible hand” was invented, we can let it work for us! The sensor networks
behind the “Internet of Things” enables us to realize Adam Smith’s brilliant vision of
self-organizing systems for the first time in human history. While this is an unprecedented opportunity to make our increasingly complex world manageable again, we
need to fundamentally change the way we think about global governance.
18 Couldn’t we emulate a decentralized system by a centralized one? This seems plausible, but
besides being more expensive, it is not obvious which data is relevant and what data obfuscates the
truth. The relevant signal in one application might be noise in another application and vice versa.
Therefore, in contrast to the conventional wisdom on Big Data, less information can be better. Let
me give an example. As I previously discussed, it is possible to understand the spread of epidemics
combining small datasets of infection cases and airplane trips. This approach performs better than
a large centralized system powered by Big Data such as Google Flu Trends. The reason is that too
much data produces problems such as “spurious correlations” and “over-fitting”, i.e. the fitting of
random, meaningless patterns. In other words, results of Big Data analytics may be irrelevant and
misleading, which can cause bad decisions.
19 Note that eliminating network effects by world markets and free trade is not a perfect solution, as
it eliminates the niches which are needed for innovation. Moreover, externalities will always play
a role when production takes place in densely populated areas or in vulnerable ecosystems.