USEMP Use cases

USEMP will demonstrate the applicability and value of its achievements through the elaboration on two use cases: (i) an OSN presence awareness and control scenario and (ii) an information monetisation scenario. The first scenario aims to empower the user with tools that easily remove or control the visibility of personal information shared on social networks. The second scenario targets the discovery of information that has economic potential in the user’s information flow and its valorisation. Both scenarios are placed in a user assistance perspective, translated into user control over automated processing and work towards explainability of obtained results. As explained in Section 1.1.1, we discriminate between three types of personal data: volunteered, observed and inferred.

Scenario 1. Developing the USEMP OSN Presence Control tool

In this scenario, the USEMP platform provides tools needed by users for improved control of the content and information that they share online and that can be observed or that can be inferred by accessing their OSN accounts. The main objective is to propose tools that allow the user to easily change the visibility for other users or the availability for the OSNs. Two groups of functionalities are offered: (a) real-time OSN presence management and (b) long term OSN presence management.

(a) Real-time OSN presence management: When they join social networks, most users associate, explicitly or implicitly, privacy levels with specific types of information they might volunteer to share. Typically, when joining OSNs, users set privacy criteria concerning their demographic data (age, gender, location) or other information they might consider private, such as relationship status, sexual orientation or political opinions. Over time, they get accustomed to the OSN and start providing data that can be used to determine ‘facts’ that were initially considered as private (i.e. not to be shared). These ‘facts’ are determined by correlating volunteered data, browsing behaviour and automated inferences. USEMP will study the interplay of such data and provide feedback about such ‘facts’. Since the ‘facts’ are based on statistical inferences and machine learning they are not facts in the ordinary sense of the term, but they may influence decisions taken with regard to the user as if they are normal facts[1]. The platform will ask the users to define the level of privacy for certain types of potentially sensitive content or behavioural data in order to make them aware whenever content or information that was volunteered, observed or inferred breaches their privacy settings. Users will be able to define in a manageable way different privacy rules to deal with different types of ‘facts’, in a way similar to that used by interfaces such as IFTTT[2]. These created rules will guide multimedia information extraction approaches to analyse new data and to correlate it with data already shared, observed or inferred, in order to raise users’ awareness of unintended private information sharing and to provide suggestions about the appropriate visibility level. In that way, users will be able to influence the –for them invisible– process that is monitoring their online behaviour.

(b) Long term OSN presence management: Users volunteer to share data which are initially harmless but whose status can change over time, taken alone or in combination with other data shared by the user or by his social entourage. USEMP will provide visualisation tools that will allow the user to have a quick view of the privacy level for personal data types they consider sensitive and to understand how personal data are exploited by OSNs to profile them. An important functionality will concern the possibility given to the user to remove/change the visibility of data that are considered sensitive. For instance, photos showing the user in certain circumstances or with given people can hurt their reputation and the user should be able to quickly change their visibility.  In order to facilitate the control of personal information, the user will be assisted by content analysis tool that filter and rank personal information based on the user’s requests for removal/visibility change. In the light of the upcoming right to be forgotten and the current right to have data erased, the user will also be given the opportunity to have personal data erased that was observed or inferred by the OSN or third parties.

The two functionalities of this scenario contribute to improved online management of personal data and to the enforcement of data protection legislation to the extent that this requires that the user is able to change the visibility and availability of information easily by influencing the invisible processing going on upon their data (this could e.g. implement the right to withdraw consent for data processing at any time, and relates to the upcoming right to be forgotten). The USEMP tool thus exemplifies an instance of Data Protection by Design.

Scenario 2. Developing the USEMP Economic Value Awareness Tool

In this scenario, the USEMP platform will provide means to contribute to raising users’ awareness concerning the economic value of their data, which is currently utilised exclusively by the data controller (OSN). Anticipating the upcoming data protection legislation, notably transparency rights in the case of measures based on profiling, the right to withdraw consent at any time, the right to have data erased and the right to data portability, USEMP offers the possibility to (a) raise the users’ awareness concerning the economic value of the personal information that they share through OSNs and (b) license for use to third parties a part of the data they share, or which have been observed or inferred. The purpose of the latter is to provide users with an even more direct experience of the value of their personal data, and in that way empower them when engaging with OSNs, while also exemplifying Data Protection by Design.

(a) Awareness of Economic Value of Personal Information: Current OSN business models are based on the monetisation of personal information volunteered by, observed of and inferred about their users. USEMP will provide an interface that will raise users’ awareness about what their personal information is worth from an economic point of view and how it is exploited by OSNs. The user will be able to understand which part of their online information is more likely to be monetised and how this process is taking place. Making use of legal and user studies, a set of multimedia information extraction techniques operating on volunteered and behavioural data will be exploited to assist the user in the proposed task. These techniques include: online behaviour analysis, multimedia content similarity to given brands, product detection in texts and images and opinion mining applied to user contributions. For instance, the user will be able to understand which content triggers specific advertisements proposed by the OSN, and which content is related to what brands. This functionality responds to a user need for more transparency concerning OSN business models and implements the legal obligation of profile transparency under the upcoming Personal Data Protection.

(b) Personal Content Licensing: in order to take complete control of the monetisation of personal information by publishers, operators, advertisers and data holders of any type, USEMP will attempt to design a framework for a licensing and control mechanism for the personal information of users, similar to the one used by intellectual property holders in other industries such as the music or film industry. Providing their specific consent in all phases of the process, USEMP will implement a functionality that gives OSN users the possibility to license personal information that they share and, at the same time, stay aware of when and to what extent their information is being used for the purpose of profit. In order to avoid the commodification of personal information and the OSN playground to become an open market for the personal data of users, USEMP will define the legal framework as well as a rewards mechanism for licenced information that will prevent this from taking place and instead provide the tools to non-profit endeavours to leverage the contributed user data, e.g. for the purpose of conducting user-centred research, all in an anonymised way. To further discourage the commodification of personal information, the value of personal data will decrease if the frequency of such data in user posts increases.

For instance, if the user shares an image depicting a brand that she likes, the platform detects the brand and the user is financially rewarded with a discount on the particular brand according to the estimated value that the information she provides has to the particular company if the presence of the brand is highlighted. Another example is that of user generated content (photos, comments) that is increasingly exploited in news materials and for which OSN users that create content are not currently rewarded and in cases not even accredited. This scenario comes as a complement to the preceding one and exploits the same underlying technical and legal capabilities. In addition, special care will be taken in order not to create an incentive structure that drives people to create personal content with economic gain as motivation. To discourage such behaviour, the value of monetized data will decrease if the frequency of such data in user posts increases.

The two functionalities of this scenario give non-commercial content creators the means to understand how their content and behavioural data is used and how they might profit from some of their data. The scenario will reinforce the collaborative dimension of business processes that actually involve OSN users but, for the moment, benefit only the OSNs and the interested stakeholders. It assumes that the current situation, whereby OSNs mine and sell personal data without granular informed consent is probably unlawful, from the perspective of data protection law. Similar questions are raised in relation to copyright protection. Endorsing these use cases and exploring the legal assumptions and implications should further clarify these issues.