Anonymity Assessment – A Universal Tool for Measuring Anonymity of Data Sets Under the GDPR with a Special Focus on Smart Robotics

Abstract

As soon as personal data is processed, European data protection law (esp. the GDPR) provides very strict rules that must be observed by data controllers and processors. This leads to a variety of problems, especially in the case of artificial intelligence (AI) systems and smart robotics, as the GDPR takes these new technologies insufficiently into account. By introducing the method of “Anonymity Assessment,” we propose an interdisciplinary approach to classifying anonymity and measuring the degree of pseudo-anonymization of a given data set in a legal and technical sense. The legal provisions of GDPR are therefore “translated” into mathematical equations. To this end, we propose two scores: the Objective Anonymity Score (OAS), which determines the risk of (re-)identifying a natural person under objective statistical measures; and the Subjective Anonymity Score (SAS), which takes into account the subjective perspective of data controllers or processors.

Keywords: anonymity, anonymization, GDPR, data protection, re-identification, Anonymity Assessment, anonymous data, PII, personal data, pseudonymous data, GDPR, Artificial Intelligence, training data, smart robotics, risk-based approach, data privacy, technical measures, anonymity score