Tackling the Myriads of Collusion Scams on YouTube Comments of Cryptocurrency Videos

Proceedings of Recent Advances in Natural Language Processing, pages 1066–1075 Varna, Sep 4–6, 2023 https://doi.org/10.26615/978-954-452-092-2_114 1066 Tackling the Myriads of Collusion Scams on YouTube Comments of Cryptocurrency Videos Sadat Shahriar University of Houston, Texas, USA sshahriar@uh.edu Arjun Mukherjee University of Houston, Texas, USA arjun@cs.uh.edu Abstract Despite repeated measures, YouTube’s com- ment section has been a fertile ground for scam- mers. With the growth of the cryptocurrency market and obscurity around it, a new form of scam, namely “Collusion Scam” has emerged as a dominant force within YouTube’s comment space. Unlike typical scams and spams, collu- sion scams employ a cunning persuasion strat- egy, using the facade of genuine social interac- tions within comment threads to create an aura of trust and success to entrap innocent users. In this research, we collect 1,174 such collusion scam threads and perform a detailed analysis, which is tailored towards the successful detec- tion of these scams. We find that utilization of the collusion dynamics can provide an ac- curacy of 96.67% and an F1-score of 93.04%. Furthermore, we demonstrate the robust predic- tive power of metadata associated with these threads and user channels, which act as com- pelling indicators of collusion scams. Finally, we show that modern LLM, like chatGPT, can effectively detect collusion scams without the need for any training. 1 Introduction The most popular online video-sharing platform YouTube has seen a surge of scams and spam com- ments since its creation in 2005. Although mea- sures have been taken, financial frauds, especially related to cryptocurrency investment have not been slowed down (Dig, Accessed: 2023-05-14). Scam- mers have adapted their tactics to circumvent the scam-detection algorithm by adopting the disguise of genuine users, engaging in seemingly ordinary conversations, and perpetrating a previously undoc- umented form of deceit known as the “Collusion Scam”. Due to their facades, such scams frequently go unnoticed by automated detection systems, pos- ing a significant threat to users who may unwit- tingly fall victim to such schemes. Consequently, it has become imperative to employ rigorous lin- guistic, psycholinguistic, and metadata analyses to effectively detect and combat these collusion scams. The “Collusion Scam” can be defined as a fake conversation where the participants pretend to be beneficiaries of a person or an entity to entrap users for their monetary gain. Typically, a scammer or a group of scammers will share their success and gratitude in working with a person or entity. Often another group joins the conversation by pretending to be curious or newbies, and on later turns they also express to be a beneficiary. In this method, the scammers share the entity’s handles or contact information to get around YouTube’s rules. Figure 1 shows an example of the collusion scam where some scammers engage in a conversation by pre- tending to be a beneficiary of a cryptocurrency investment through a claim expert. The rise of cryptocurrencies has not only at- tracted genuine enthusiasts and investors but has also unfortunately attracted a surge in fraudulent activities. The absence of comprehensive …

Content truncated to 500 words