Scoring algorithms are increasingly deciding whether consumers can order on account or take out insurance. This harbors the dangers of errors and discrimination.
ACar insurance, loans, online payments: algorithms are being used in more and more areas of life to classify consumers. According to advisers to the federal government, however, so-called scoring harbors the risk of errors and discrimination. For example, it is critical if the creditworthiness is made dependent on someone living in a good residential area, said the Advisory Council on Consumer Affairs on Wednesday.
Consumer Minister Katarina Barley (SPD) emphasized: “Any form of scoring must be transparent and rule out discrimination and abuse.” Independent bodies would have to check the calculations and needed more support for this.
Scoring algorithms decide, among other things, who can buy on credit on the Internet and who can only order in advance. Health and auto insurance terms and conditions can also change depending on your personal score. Consumers would have to be informed about their evaluation free of charge, called the Expert Council.
Barley ruled out a social point system, which will be used in China from 2020 to differentiate between “good” and “bad” citizens. “A scoring system like the one the Chinese government is striving for and is currently testing is incompatible with our legal and value system,” she said. In China, points are deducted, for example, for traffic offenses or late payment on invoices, and the account can be topped up with donations or volunteer work.