The field of social media research is moving towards a deeper understanding of online discourse and its implications on society. Recent studies have focused on the collection and analysis of multimodal social media data, including the comparison of study-created and genuine content. This research has shown that study-created data can be valuable for training models, but may differ from genuine data in significant ways. Additionally, there is a growing interest in the use of machine learning methods to enhance online deliberation and detect violent content. The fragmentation of social media spaces is also being explored, with studies investigating the characteristics of political discourse and polarization on emerging platforms. Furthermore, researchers are working to improve regulatory oversight in online content moderation and provide more accurate data for decision-making. Noteworthy papers include:

• A study on cross-platform violence detection, which introduced a dataset of 30,000 posts hand-coded for violent threats and achieved high classification accuracy.
• A survey on natural language processing to enhance deliberation in political online discussions, which showcased the potential of machine learning methods to counteract issues in online discussions.
• A case study on the use of online forums for legal information, which found that crowdsourced legal information tends to be legally sound but sometimes incomplete.