This panel examines remarkable inferences from personal data: the predictability of human behavior by personal data. Personal data is often used via profiling or scoring techniques, which have been an integral part of our society for decades, to predict the future of an individual. However, the possibilities of profiling and scoring have reached a new quality. This panel explores the worst-case scenarios of algorithmic discrimination (e.g. mortgage rejections). On the other hand, the panel shares good practices by the risk-based approach in order to minimize the risks for rights and freedoms of data subjects (e.g. companies starting to effectively communicate with their potential costumers). The panel also introduces Japanese and Australian cases in comparison with the European context.
• How can we make profiling decent in terms of the risk-based approach?
• Are there gaps in profiling regulations we should mind in theory and in practice?
• What are the implications of the information bank plan in Japan and the facial-matching system and Robodebt disaster in Australia?
• Does profiling design for the AI and robotics in a privacy friendly manner?