Impact of the Digital Services Act: A Facebook Case Study
Impact of the Digital Services Act: A Facebook Case Study

Author(s): Tetiana Haiduchyk, Marat Khusnutdinov, Fabio Belafatti, Jakub Kubś, Andrius Šuminas, Gundars Bergmanis-Korats
Subject(s): Law, Constitution, Jurisprudence, Media studies, Communication studies, Theory of Communication, ICT Information and Communications Technologies
Published by: NATO Strategic Communications Centre of Excellence
Keywords: Digital Services Act; Facebook; Case Study; Digital Services Act; social media;
Summary/Abstract: The adoption of the EU Digital Services Act (hereinafter the DSA) aimed to create a safe online information environment, specifically within the EU/EEA area. The aim of this research was to measure the effects of the DSA in curbing the spread of harmful content on social media. As measuring the results of such a broad goal was challenging, our study focused on one of the dominant social media platforms: Facebook. To assess the impact of the DSA, we compared the share of harmful content published by Polish and Lithuanian accounts on Facebook before and after the DSA entered into force. Our multi-stage approach involved using a small AI model, GPT-4o mini, to initially flag harmful content, followed by applying larger models for validating and in-depth reasoning. In total we classified 959 harmful posts from 2023 and 1,392 posts from 2024. Hate speech targeting individuals based on protected characteristics remains the platform’s most significant vulnerability in combating harmful content. About 90% of such content for both languages was related to hate speech in both 2023 and 2024. However, differences in the numbers of posts with harmful content were noted between the two languages. In 2024 such posts in Polish increased by 55%, while those in Lithuanian declined by 11%. Differences between the languages were also observed in the topics of harmful content, which may be influenced by varying model performance across the languages. However, it should be noted that the study transparently shows the accuracy metric in the Appendix. While antisemitism mostly related to the Israel–Hamas war was the most prevalent in both languages, the number of harmful messages related to this topic increased only in Polish in 2024. Therefore, we could argue that it is crucial for platforms to dedicate more efforts to monitoring and responding to the content associated with ongoing political conflicts, since those are essentially a fertile ground for further hostilities online. At the same time, the increase in harmful posts varied significantly by account type in Polish. Specifically, individual Polish Facebook accounts published 6% more harmful posts, while Polish groups saw an increase of 128% in these posts in 2024. Such a noticeable difference likely indicates the platform’s system vulnerabilities in detecting and addressing harmful content within groups. An assessment of the platform’s efforts to ensure a safe information environment before and after the implementation of the DSA showed dual results. On the one hand, we noted that the share of fact-checked posts increased significantly in 2024, likely suggesting an additional platform investment in this area. On the other hand, we tracked the decline in the removal rate of harmful posts, which was particularly sharp for Lithuanian content. This may indicate insufficient efforts to monitor and remove violations involving small languages. As a result, the study demonstrated that despite certain improvements the platform made in creating a safe online environment, we could not claim an overall enhancement after DSA enforcement. Additionally, our results highlight the current vulnerabilities and areas for improvement that the platform should address. This research is just a glimpse into a field of research with an expansive scope. Therefore, our conclusions and recommendations should be seen as an encouragement for further and wider research across online social media platforms.

• Details
• Contents