Field studies with autonomous robots and data protection

ZEN-MRI aims researching the use of autonomous robots in public spaces. In order to obtain meaningful results on the behavior and attitudes of people in relation to autonomously operating machines–such as driverless cleaning devices– field studies at real locations under real conditions are essential. This not only requires special demands on research design and practical realization. Sufficient attention must also be taken to data protection.

Conducting studies under real conditions with legal certainty

Especially in field studies at real  locations with persons who have no special experience with robots, it is often unclear to scientists and researchers how to proceed from a legal perspective. A common complaint is: Complex laws and legal frameworks with sometimes opaque or incomplete provisions may slow down research under real conditions. ZEN-MRI counteracts this by carrying out legal monitoring right from the beginning and throughout the whole project.

And indeed, the data protection obstacles in field studies with autonomous robots are not insurmountable. Above all, it is important to consider the data protection requirements right from the start of the research design process and to  consistently implement all necessary steps during the realization. Ideally, the legal framework should be designed in such a way that researchers can act as freely as possible and at the same time to operate–for example towards randomly involved passers-by–transparently and legally.

A smart data protection concept for legally compliant research

The necessary basic framework of every behavioral study should be a data protection concept. Such concept has to address the study design as well as questions like the research purposes persued, legal responsibility, type and manner of data collected, and last but not least, the legal basis and the duration of processing and subsequent deletion of data. A decision must also be made on ways to pseudonymize or anonymize research data. When integrating special software, AI systems, research service providers or cloud services, all associated data protection issues must also be clarified.  

A particular challenge in interaction studies under real conditions: As a rule, the purpose of this research is to obtain results that are as unaffected as possible by external environmental influences. For example, as part of field research in the ZEN-MRI project, passers-by should be as unbiased as possible when encountering the autonomous robots. The difficulty therefore lies in designing the experimental set up legally in such a way that the affected persons are treated transparently and informed, while at the same time the purpose of the research is not jeopardized.

Research privilege – data collection in the interests of science

The research privilege under data protection law is a great help here, as it allows to take into account the principles of freedom of science and research even in face of data protection laws such as the GDPR. It allows universities and research institutions to carry out their research mandate without disregarding data protection, but nevertheless appropriately to consider the research purposes pursued. For example, the research privilege–with a fundamentally transparent, data sensitive and secure approach–can restricting data protection processes in favor of uninfluenced human interaction behavior. That means, the research privilege allows to carry out data collection under slightly less strict requirements for the benefit of research.

However, research privilege also has its limits and must always be brought into line with the data protection interests of affected persons for every study conducted: The more sensitive, comprehensive and long-term the observational and evaluation studies are, the sooner the usual data protection standards apply. To conclude, the research privilege cannot justify every research interest in human behavior.

Someone is always responsible

Even if the research privilege under data protection law offers some freedoms in the design and implementation of human interaction studies, it does not–for example– generally exempt from data protection concepts, records of processing data collections, data protection information and technical and protective measures (TOMs). Despite research privilege, there is still a lot to consider–depending on the research design and purpose. Proper legal implementation is insofar not just a burdensome duty. Above all, it gives researchers legal certainty, as they and their research institutions are in the end responsible for their studies and data processing procedures.


Further sources:

– https://www.netzwerk-datenschutzexpertise.de/sites/default/files/gut_2018_forschungklauseln_181018.pdf