At re:publica 2017, Strategy Of ‘The Facebook Empire’ Revealed By Patents

In a talk at the re:publica 2017 in Berlin this week, academics from the Share Lab Project presented how they relied on an unusual resource to get a measure of the algorithms of Facebook. By reading through a part of the 8000 patents registered by the company, the researchers were able to shed some light into the process of how the “Empire” turns the raw data they treat their users as into those valuable big data golden profiles that then can be marketed.

Patents like “Systems and Methods for Social Mapping”, “Social Data Recording,” and many others allowed Prof. Vladan Joler, director of the Share Foundation and chair of New Media Department at University of Novi Sad, Serbia, to map three different stores that feed the social graph, which according to Joler is the very heart of the Facebook empire. In essence, it is the way in which every piece of information derived from the user or additional external data sources is set in relation in profiling the user.

The three stores referenced in the patents were the “action” store covering users’ actions, the “content” store holding objects of various types of content, and the “edge” store that comprised information between users and other objects.

In the panel discussion that was part of the Media Convention subconference of the re:publica, experts considered what should or could be done about the growing influence of Facebook over public opinion, with the number of users coming close to 2 billion and the platform equalling “the Internet” in parts of the world.

One path could be an obligation for neutrality to allow users themselves to decide how their newsfeed would be constructed and from which sources, according to the experts. A better understanding of the layers of algorithmic machines is indispensable, according to Joler, because hidden inside the black box are human rights violations, new forms of exploitation, as well as mechanisms of manipulation to influence billions of people.

The danger of algorithms, better ways to conquer and shape algorithms as well as the question of fake news, manipulation of leaks and the potential “hack” of elections and democracies are all topics at this year’s 11th re:publica.

The conference that has the motto “love out loud” this year has become the biggest media conference with 1180 speakers over three days. The organisers during the opening session warned against right-wing populism and the degradation of freedom of the media.

The opening keynotes were testimony for the bad news. Hungarian journalist Márton Gergely described the hostile takeover of the political daily Népszabadság by the government in 2016. Katarzyna Szymielewicz, president of Panoptykon Foundation in Poland, reported about how commenting on new legislation in Poland had gone from being safe to being so dangerous that she and her team members considered if the time would come to leave the country.

The Berlin audience gave standing ovations to Can Dundar, former chief editor of Turkish daily Cumhuriyet who was prosecuted in Turkey for breaking the story on Turkey’s provision of weapons to ISIL fighters and lives now in exile in Germany.

A less friendly reception could await German Minister of the Interior Lothar de Maiziere who is announced to talk about net politics on 10 May. De Maiziere might face tough questions over a recent draft to strike on fake news, as well as new legislation allowing for state hacks against suspects and more competencies for intelligence services.

As in the year before, the topic of refugees and integration also is on the agenda, as well as the flurry of reviews of major European Union legislation touching the net from the e-privacy to the copyright directive.

 

Image Credits: MCB / Uwe Völkner