A Review of Techniques for Detecting Illicit Messages on Twitter

Twitter is a very broad social network, allowing people to communicate with each other and express their ideas, thanks to its short and quick approach to posting. Unfortunately, it is not exempt from illicit affairs occurring on the platform. One problem in social networks is how they promote and spread illegal services, such as human trafficking, prostitution, drugs, etc. Since the rise of Machine Learning (ML) and Deep Learning (DL) solutions, different approaches and algorithms have been developed and implemented to avoid those issues. Thus, it is important to know about AI advancements to analyze the effectiveness and need for possible future solutions. In this work, we have compiled previous research on detecting illicit messages on Twitter using ML and DL techniques. We found that ML approaches are still popular and useful, but deep learning-based approaches, mainly Transformer-based ones, are better and more adaptable to different tasks and contexts.