100
Crowdsourcing Technology: Definitions and Evolution
In the context of such constraints in our current surgical training paradigm, increas-
ing interest in crowdsourcing has developed. This term refers to a large-scale
approach to accomplishing a task by opening it to a broad population of decentral-
ized individuals who may complete the task more effectively in aggregate than any
single individual or group of individuals [ 31 ]. Despite its recent rise in popularity,
such a large-scale approach to problem-solving dates back as far as 1714, with the
British Parliament’s establishment of the Longitude Act. This declaration offered up
to £20,000 to any individual who could provide a simple, practical, and accurate
method of determining the longitude of a ship at sea. It has since been applied
across multiple fields from astronomy to ornithology, with the advantages of effi-
ciency, scalability, flexibility, and diversity to solving the particular problem at hand
[ 32 ]. The success of crowdsourcing lies in aggregating the collective intelligence of
all participants, such that the distributed wisdom of the group surpasses that of any
single individual [ 33 ].
Over the past few decades, the widespread integration of the Internet has further
helped to shape the evolution of crowd-based wisdom. The aggregate value of
crowds networked through the Internet has been termed by some as “collective
intelligence,” an entity that is constantly changing, growing, and evolving in real
time [ 34 ]. The Internet has not only allowed for the rapid connection of individuals
in the pursuit of a single problem’s solution but has also facilitated the formal orga-
nization of such virtual individuals into an easily accessible entity that can solve a
wide range of problems, from straightforward tasks to complex problems requiring
critical intellectual input.
The Amazon Mechanical Turk is one example of this phenomenon. This market-
place service provides access to more than 500,000 “Turker” crowdworkers from
over 190 countries who perform a range of “human intelligence tasks,” including
data processing, information categorization, business feedback, and content mod-
eration. Such tasks are generally deemed too challenging or inefficient for current
artificial intelligence algorithms to complete. For each completed task, workers are
paid a small sum. One survey of these workers actually found that monetary incen-
tive was the primary reason for their participation in such tasks, particularly for
those not based in the United States [ 35 ].
Such organization of crowdworkers has allowed for the recent rapid and wide-
spread adoption of crowdsourcing across a multitude of fields. Within the business
realm, several corporations have been built around crowd-based wisdom, capitaliz-
ing on their collective intelligence to sell merchandise, amass repositories of photo-
graphs, and develop innovative research and development solutions [ 33 ]. The reach
of crowdsourcing also extends into the healthcare realm, where it has served a myr-
iad of roles in the areas of molecular biology, comparative genetics, pathology, and
epidemiology [ 36 ]. Challenges in computational molecular biology have been read-
ily addressed by large-scale problem-solving, which has been used to discover ter-
tiary protein folding patterns [ 37 ] and to generate phylogenetic arrangements of
genetic promoter regions from vast numbers of nucleotides [ 38 ]. Crowdsourcing
J.C. Dai and M.D. Sorensen