210
A. Solanas et al.
12.3.5 Intelligence Privacy
In the current globalization context, there are numerous service providers offering
similar products and services, which results in an increase in the number of
competitors. Data collected by each provider is, in many cases, very valuable, and it
is used to extract knowledge and provide customer-oriented, personalized or addedvalue services. Hence, sharing and releasing this data is not a common practice,
especially if competitors have a chance to take advantage from such data. However,
in some cases, organizations (not necessarily competitors) could take mutual benefit
by collaborating, but they do not want to share their data. This situation is covered
by what we have called intelligence privacy. In this dimension, the goal is to allow
the collaboration among several organizations so that all could make joint queries
to databases to obtain joint information in such a way that only the results are
revealed (i.e., the actual information in the databases of each company is not shared
or revealed).
To clarify the concept of intelligence privacy let us have a look at the following
example of manufacturers of autonomous and intelligent vehicles. Each vehicle
manufacturer integrates many built-in sensors on vehicles to gather, store and
analyze the status of the car, the nearby environment and further driving-related
parameters. Since these data are highly sensitive, manufacturers might decide not to
share them. However, collaboration among manufacturers by sharing data could be
extremely useful to improve safety on roads and to avoid collisions. In this sense,
each manufacturer (even if they compete) would benefit from the collaboration, that
is, to obtain joint results, but they want to avoid sharing their intelligence data.
In this situation of mutual distrust, Privacy-Preserving Data Mining (PPDM)
techniques emerge as the natural solution to protect intelligence privacy [12]. PPDM
methods are applicable once independent entities want to collaborate to obtain
common results that benefit both of them, but without sharing their data since they
do not trust each other. In such scenario, by applying PPDM to the queries submitted
across several organization databases, the amount of information transferred to
every party is controlled, and this does not pose risks that original data will be
revealed, only the results.
It is worth emphasizing that Intelligence Privacy considers data belonging to
companies (e.g., the heat dissipated by the front-left wheel brake). Thus, data
collected by companies but belonging to individuals should not be considered under
this dimension because they belong to the users and not to the companies, and hence
they should be managed as such.
12.4 Future Trends and Challenges
Privacy has often been considered from a data-oriented perspective. The goal was to
protect the data regardless of their origin. Data, in this sense, were seen as something
of value that belong to whoever has the ability to collect, store and exploit them and,
A. Solanas et al.
12.3.5 Intelligence Privacy
In the current globalization context, there are numerous service providers offering
similar products and services, which results in an increase in the number of
competitors. Data collected by each provider is, in many cases, very valuable, and it
is used to extract knowledge and provide customer-oriented, personalized or addedvalue services. Hence, sharing and releasing this data is not a common practice,
especially if competitors have a chance to take advantage from such data. However,
in some cases, organizations (not necessarily competitors) could take mutual benefit
by collaborating, but they do not want to share their data. This situation is covered
by what we have called intelligence privacy. In this dimension, the goal is to allow
the collaboration among several organizations so that all could make joint queries
to databases to obtain joint information in such a way that only the results are
revealed (i.e., the actual information in the databases of each company is not shared
or revealed).
To clarify the concept of intelligence privacy let us have a look at the following
example of manufacturers of autonomous and intelligent vehicles. Each vehicle
manufacturer integrates many built-in sensors on vehicles to gather, store and
analyze the status of the car, the nearby environment and further driving-related
parameters. Since these data are highly sensitive, manufacturers might decide not to
share them. However, collaboration among manufacturers by sharing data could be
extremely useful to improve safety on roads and to avoid collisions. In this sense,
each manufacturer (even if they compete) would benefit from the collaboration, that
is, to obtain joint results, but they want to avoid sharing their intelligence data.
In this situation of mutual distrust, Privacy-Preserving Data Mining (PPDM)
techniques emerge as the natural solution to protect intelligence privacy [12]. PPDM
methods are applicable once independent entities want to collaborate to obtain
common results that benefit both of them, but without sharing their data since they
do not trust each other. In such scenario, by applying PPDM to the queries submitted
across several organization databases, the amount of information transferred to
every party is controlled, and this does not pose risks that original data will be
revealed, only the results.
It is worth emphasizing that Intelligence Privacy considers data belonging to
companies (e.g., the heat dissipated by the front-left wheel brake). Thus, data
collected by companies but belonging to individuals should not be considered under
this dimension because they belong to the users and not to the companies, and hence
they should be managed as such.
12.4 Future Trends and Challenges
Privacy has often been considered from a data-oriented perspective. The goal was to
protect the data regardless of their origin. Data, in this sense, were seen as something
of value that belong to whoever has the ability to collect, store and exploit them and,