Ο μισαλλόδοξος λόγος στο διαδίκτυο και ζητήματα ως προς την ανίχνευση μισαλλόδοξου λόγου
Hate speech on the internet and challenges in its detection

View/ Open
Keywords
Μισαλλόδοξος λόγος ; Ρητορική μίσους ; Ελευθερία έκφρασης ; Hate speech ; Διαδίκτυο ; Μέσα κοινωνικής δικτύωσης ; Ν. 4285/2014 ; Απόφαση-Πλαίσιο 2008/913/ΔΕΥAbstract
This thesis examines the phenomenon of hate speech on the internet, focusing on the legal and technological challenges arising in its detection and response. A central issue of concern is the balance between the constitutionally enshrined freedom of expression and the obligation to protect human dignity from manifestations of online hatred. The methodology is based on bibliographic research, combined with an analysis of technological tools for identifying hate speech. The first chapter analyses freedom of expression through the lens of national, EU, and international law, with emphasis on the limitations that may be imposed when it conflicts with manifestations of intolerance. Special reference is made to the specific features of the online environment and the role of algorithms in the flow of information. The second chapter focuses on the conceptual and legal definition of hate speech, highlighting the difficulty of establishing a uniform definition and presenting the interpretative approach of the ECtHR, primarily in light of Articles 10 and 17 ECHR. In this context, the chapter explores its digital dimension and the role of platforms that foster its dissemination. The third chapter presents the international, European, and national regulatory framework, with emphasis on the case law of the Hellenic Supreme Court, while the fourth chapter analyses the technological solutions employed in detecting hate content, particularly through artificial intelligence and machine learning, highlighting inherent shortcomings such as limited accuracy and the inability to assess context. The study’s findings reveal both the interpretative and practical gaps in the current legal framework and the limited capabilities of technological tools. The thesis concludes with policy recommendations, including the combined use of legal, technological, and educational instruments, the enhancement of algorithmic transparency, and the promotion of interoperable mechanisms of cooperation between public and private actors, with respect for fundamental rights. This interdisciplinary approach is deemed necessary for the effective combat of the phenomenon, in line with contemporary academic and regulatory standards.