Page 110 - ITU Journal - ICT Discoveries - Volume 1, No. 2, December 2018 - Second special issue on Data for Good
P. 110
ITU JOURNAL: ICT Discoveries, Vol. 1(2), December 2018
benefits of data-restrictive choices and data sharing information.
respectively.
In the end, data gathering and processing
Still, individual caution and data-restrictive choices organizations are the entities who develop and
only lead so far. It is up to regulators and data implement innovations, and who determine the
processing organizations to ensure that easy-to- extent to which users can be data sovereigns.
understand information on data-intensive ICT is Service providers can support this process through
available and disseminated. In particular, careful technological infrastructures that allow the
reflection is needed on how the ideal of individual to control the flow of her data. In the ideal
transparency could be attained. Transparency can case, data processing rests on the informed consent
imply disclosure. However, requirements to of the individual whose data is being processed,
disclose technological designs, code, and data while she retains options to withdraw consent and
processing algorithms raise several difficulties. mandate deletion of her data from the service
First, even if disclosed, such information remains provider. Controllability would go a long way
unintelligible to the lay user, and thus does not towards harmonizing the benefits of
advance her informedness. To some extent, terms big-data-driven de- and re-contextualization with
and conditions suffer from this problem as well: the privacy and expectations of individual data
everybody accepts them, almost nobody reads them, subjects.
and only some of those who read them understand
them. Disclosure of code, just as terms and Governments can take on a key role in this process,
conditions, would have to be complemented by by encouraging the self-regulation of organizations.
societal discourses with a variety of stakeholders For example, independent, industry-wide data audit
and experts. Only then can the disclosure of and certification centers can make responsible data
complex technical contents end up guiding the non- management visible. Where self-regulation is
expert user. Second, even partial mandates to lacking, the state can take over through regulation,
disclose codes and algorithms can affect business monitoring, and sanctions. The GDPR is the most
interests and incentives for innovation on the side recent example of the range of instruments that can
of the ICT providers, e.g., if this de facto be employed internationally while leaving
compromises protections of IP. A more promising individual nation states discretion to spell out the
avenue is to mandate disclosure of the purposes and precise nature of these tools in their jurisdiction.
aims of a given algorithmic tool. This would make it
possible for outsiders to get an idea of the intended The rollout and alignment of technological
functioning of the tool, and to assess whether it standards for data interoperability and
works as advertised, e.g., reaches the intended goal programmatic interfaces is an important area
with the proclaimed precision and without undue where industry and policymakers can work
discrimination against certain populations [4,5]. together to harness data. Uniform standards and
formats for exchanging and connecting data from a
Moreover, disclosure by itself is unhelpful if variety of sources and between different systems
individuals lack room for maneuver. Amongst the make data comparable and translatable. It also
features of the platform economy is that its players facilitates quality control and documentation.
benefit from economies of scale and network effects
which sooner or later lead to market concentrations Data interoperability and tools to link, organize,
[10]. This effectively constrains the choices of filter, and curate data efficiently [39] can yield
individuals for moving from one platform to significant benefits towards the normative ideal of
another. As inforgs, it is out of the question for data sovereignty. First, interoperability is key to
individuals to refrain from using ICT services. One data sharing. Lack of interoperability does not
condition of data sovereignty is thus that policy and necessarily threaten privacy, but it does
lawmakers find ways to uphold competition and compromise potentials for exchanging data, a
ensure that the market offers a plurality of challenge for example in endeavors to utilize
data-intensive services from providers with routine clinical data for medical research [40].
different privacy and control mixes. This also Second, while standardization does not by itself
involves discourses on the pricing of data [38] in advance data sovereignty, it sets the stage for
order to compensate individuals for value introducing technological solutions that help
generated through the processing of their individuals to control the flow of their data.
88 © International Telecommunication Union, 2018