Anonymization for the GDPR in the Context of Citizen and Customer Relationship Management and NLP

Gil Francopoulo , Léon-Paul Schaub (1)

1 LIMSI - Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur

The General Data Protection Regulation (GDPR) is the regulation in the European Economic Area (EEA) law on data protection and privacy for all citizens. There is a dilemma between sharing data and their subjects' confidentiality to respect GDPR in the commercial, legal and administrative sectors of activity. Moreover, the case of text data poses an additional difficulty: suppressing the personal information without deteriorating the semantic argumentation expressed in the text in order to apply a subsequent process like a thematic detection, an opinion mining or a chatbot. We listed five functional requirements for an anonymization process but we faced some difficulties to implement a solution that fully meets these requirements. Finally, and following an engineering approach, we propose a practical compromise which currently satisfies our users and could also be applied to other sectors like the medical or financial ones.

Mots clés: anonymization; pseudonymization; GDPR; NLP

Domaines: Informatique; Linguistique; Informatique et langage [cs.CL]; Traitement du texte et du document

Gil Francopoulo, Léon-Paul Schaub. Anonymization for the GDPR in the Context of Citizen and Customer Relationship Management and NLP. workshop on Legal and Ethical Issues (Legal2020), LREC2020, May 2020, Marseille, France. pp.9-14. ⟨hal-02939437⟩