It’s wild that the reviews run the gamut, from strong reject, borderline to strong accept . One reviewer recommends the work for a methodology award and the other thinks there’s no usefulness in the modeling approach 🫣 Any way, here is the abstract. Looking forward to present it at IC2S2 2024 & get feedback. See you there!

The abstract:

The Digital Services Act (DSA) was introduced in 2022 as a unifying framework for social media content moderation across the European Union. Although the DSA mandates notice-and-action procedures, it lacks precise requirements for turnaround time and takedown deadlines, only requiring decisions to be made in a ``timely” manner. While it is crucial to understand the impact of every regulation specification, it is challenging to evaluate content moderation regulations due to limited data access and a lack of prediction tools. We quantify the impact of takedown deadlines on the spread of unlawful content via an agent-based model informed by empirical data from the DSA Transparency Database (DSA-TDB). We find that timely response to complaints is critical in mitigating the harms caused by unlawful content, especially within the first 24 hours of a message being created. Our method provides a traceable, and replicable framework to craft policy recommendations that effectively mitigate the harms caused by abusive social media actors. The method can be adapted to accommodate the designs of different platforms and diverse country-specific regulatory contexts.