Philipp Judge about legal dangers of a big-data election campaign
The SPD plans to be advised in the future of Jim Messina, who helped 2012 as a campaign manager the Democrats around Barack Obama with Big Data Use to victory. Such election campaign methods but in Germany on legal limits. Telepolis spoke about the legal scientist Philipp Richter from the University of Kassel.
Mr. Richter, which benefits were connected for the parties with a big data election campaign?
Philipp Richter: The US Prassidal election campaign 2012 has produced a new strategy to collect and evaluate the large amounts of data on the individual eligible entitlements to address and convince specifically individuals. The data has been collected from Wahler registries, earlier election campaigns, purchased databases of commercial data collectors, from telephone surveys, social networks and the evaluation of internet usage. Thus, the analysts forecasted the election behavior of the individual [SIC] Wahler for the upcoming and for the upcoming choice. The forecasts for the reasonable choice were adjusted with the actual election result to increase the accuracy. Due to field trials, one analyzed which types of voters could be animated to elect and to elect their own candidates.
Jim Messina. Photo: Weibe house. How did that matter?
Philipp Richter: There were well prepared election cheaters targeted to the individual "wear out" sent home, they called on or contacted you about social networks. For this, if possible friends or acquaintances of the "wear out" recruited from whom you knew you belonged to your own warehouse. E-mails were adapted for specific elective types based on their personal properties. The official websites of the campaign campaigns evaluated surfing behavior and transformed this in individually appropriate electoral messages. Through the analyzes it was possible to effectively influence individuals. The election results, which were calculated on the basis of this strategy, on election day partly almost exactly one, especially in the undecided, has been achieved according to its own statements a high success rate.
"The survey will be inadvertent"
Which data protection regulations speak in Germany against such a campaign?
Philipp Richter: First, there is an organizational difference. In the US, you have to register if you liked elections. The registers can use the parties free. If you cancel them with the population, (for this purpose, for example, the customer registers of local telephone providers were used), you are namely, who was not election. In Germany, Burger does not register individually. There are directories of voters. Access to these is only very crimped. In addition, you can not keep close to the non-balk with you. It was possible with data from social networks et cetera probability calculations, but this secure knowledge, who has not chosen, did not have. You could then not synchronize with such accuracy the result of forking elections with individual electoral forecasts. A gross proportion of non-data surveys and analyzes in contradiction to German data protection law. Only one example: in Germany one may collect, process and use data, for example, in social networks, as far as non-protective interests of those affected persons own interests obviously overweight. However, from these data to predict new information about those affected, this permit is no longer included. If the purpose of the data collection is a comprehensive profile formation about political attitudes and argumentative errors, the survey will already be inadvertent.
"Clear incredit against the data protection law"
That says that that can not occur in Germany?
Philipp Richter: Unfortunately not. For example, in Germany, the like-buttons behavior of such people who released their political orientation freely compared to people who do not do that and thus forecast their political orientation. The goods, if it is not completed completely anonymous and remains, or is a consent from each person concerned, a clear incredit against data protection law. If from the data of those who are voluntarily revealing them, and which one may increase under circumstances, forecasts on them or others may be prepared, the informational self-determination against themselves. Unfortunately, this means that with Big Data’s own information behavior, especially in the political field, can be a significant risk for others.