A research project seeking to find out whether the Internet can be mined to inform credit checks has been called off due to public outrage. But the automatic collection of data remains a major trend in the IT-branch.
In Germany, nothing works without the organization Schufa - that is unless you want to pay for everything with cash. But whoever wants to take out a loan, sign a mobile phone contract, or rent an apartment has to agree to a Schufa inquiry.
That means that the bank, the cell phone service provider, or the landlord asks the credit agency whether the customer can pay the relevant interest rate or rent - whether or not he or she is "credit worthy." In order to answer that question, Schufa collects and saves data - around 514 million entries from 66 million people.
How the rating works
Schufa collects information about bank accounts, loans, credit cards, leasing agreements, as well as mobile phone and mail order accounts of people in Germany. It does not collect information about your profession, salary or nationality. If a person has debt, that's not viewed as a bad thing - debt and credit cards are considered positive by the credit agency.
But whoever is unable to meet their installment payments on time or balance their accounts receives a negative rating. Out of all the individual information collected, Schufa makes a number called the "score value." That's the number that banks, companies or landlords receive when they ask for background checks on a customer.
But the credit agency will not reveal how exactly the "score value" is calculated. Schufa has tried for some time to at least make the criteria of their evaluation more transparent. Consumer protection agencies object to some of the criteria, because the person being evaluated has little control and they can be very speculative in their interpretation.
"Geoscoring" is sometimes used as criteria - that means whether you live in a rich or poor neighborhood. And if your neighbors don't pay their bills, it can have a negative impact on your own rating. Schufa says that its does not use geoscoring, but other credit agencies are less scrupulous in that regard.
Combing the Internet
In principle, virtually all information about a person is of interest to the agency, not just a place of residence. Where you shop, what your hobbies are and were you go on vacation - that all could be useful information. But very few citizens want to be observed, scrutinized and evaluated in this way.
That's why news of a research project, which Schufa had contracted to the Hasso Plattner Institute at the University of Potsdam, triggered such a storm of outrage. The scientists wanted to find out whether and how one could cull publicly available, useful information from the Internet. The whole thing was envisaged as fundamental research in which Schufa - out of its own curiosity - asked the question: "What consequences does the technological development of the Internet have for the individual's economic existence?"
Facebook as a source?
In the public perception, the whole thing had a different effect. Did Schufa want to comb through personal comments in the Internet, for example in Facebook, in order to find hints about a person's credit worthiness?
But the project was never about that, according to software specialist Felix Neumann, who was supposed to lead the research work. From a technical perspective, it's hardly possible to draw conclusions about the creditworthiness of an individual person based on their Facebook posts, Neumann said. Schufa has high quality data anyway and does not need to access "such an insecure and difficult medium like the Web."
Information about companies, or for companies, can be better won from the Internet through automated processes, according to Neumann. What one can really read from social networks or Twitter posts are mass opinions about a particular product. "I can't effectively find out what Mr. Müller thinks about a product," Neumann said.
"Data mining," or combing through massive troves of information in the search for trends or other important information, is at the moment the number one interest in the computer science community. All over the world, research groups at universities are working on this issue. But at the Hasso Plattner Institute the project was called off three days after the first press release, due to "the many misunderstandings in the public about the agreed upon research methodology." The institute said on Friday that as a consequence of the negative public reaction the research could not continue with the peace and quiet required.
"What angers me is that we're now leaving these research questions to private companies and foreign countries because a matter-of-fact discussion is apparently not possible in Germany," Neumann said. He criticizes the press coverage which focused above all on an internal document from the institute.
"The papers which ended up in public were simply project ideas, or the results of brainstorming," said Neumann, adding that there was never a concrete project plan.
In September, the group would have published its first results - nothing will come of that now. So Schufa will have to find other answers to the question of whether information mining from the Internet will threaten its own business model some day. What's clear is that the trails that Internet users leave behind are being combed and collated at great expense, by those who run social networks - such as Google and Facebook - as well as other interested parties and groups, such as intelligence agencies.
Author: Michael Gessat / slk
Editor: Gregg Benzow