piwik-script

Deutsch Intern
Robotics Law Research Centre

Data protection

Very important are currently issues of Data protection. There are increasingly acute conflicts between issues related to privacy, security and convenience.

Semi-automated systems (such as those installed in motor vehicles) are equipped with numerous sensors to prevent accidents, and therefore damage from ever occurring. The sensors record environmental data, and then process and store the data obtained. This may also include personal data, e.g. information about the working behaviour of employees, their break times and their movement patterns, etc. This makes employee data protection issues relevant that have not yet been fully clarified. Even semi-autonomous cars can record and store a vast amount of information about their occupants, but also about other people (e.g. the occupants of other vehicles).

The principles of European data protection law stipulate that personal data may only be recorded, processed and stored if the person concerned consents or if the law explicitly permits the recording, storage or processing of the data. In addition, the principles of data economy - no more data should be collected, processed and stored than is absolutely necessary - and purpose limitation apply: Data may only be used for the purposes for which it was collected.

Compliance with the principles outlined above becomes a problem, when semi-autonomous systems recording the data are linked to the Internet. Current developments, such as those discussed under the heading "Big Data", make traditional data protection law appear completely obsolete. “Big Data" is based on the idea of deriving forecasts about the behaviour of groups or individuals from data volumes that are supposed to be as comprehensive as possible (i.e. as large as possible) by means of appropriate analyses. This is in direct conflict with the principles of data economy and purpose limitation. As soon as it becomes possible to formulate even reasonably sound predictions about the behaviour of individuals from large volumes of data, even the concept of "personal data" itself becomes dubious - potentially every data point, every computer-readable and computer-analysable piece of information, can be related to specific individuals and thus becomes "personal".