Understanding and Characterizing Cryptocurrency Free Giveaway and Arbitrage Bot Scams In the Wild

This paper presents a large-scale analysis of two prevalent cryptocurrency scams disseminated through Twitter and YouTube. The first scam involves free giveaway schemes where scammers publish fake giveaway websites to deceive victims and steal funds. The second scam revolves around arbitrage bots and publishes videos to entice victims into executing malicious smart contracts. To collect and analyze these scams in the wild, we developed a fully automated scam detection system called \textit{CryptoScamHunter}, which collects data from Twitter and YouTube and employs Nature-Language-Processing (NLP) models to automatically identify scams and extract the associated cryptocurrency address. By deploying \textit{CryptoScamHunter} over 11 months spanning from June 2022 to May 2023, we detected 95,111 free giveaway scam lists on Twitter and 10,442 arbitrage bot scam videos on YouTube that were disseminated through thousands of social network accounts and have reached millions of users. Through analysis of the scam creator accounts, we discovered that the scammers combined different strategies to spread each scam, including compromising popular accounts and registering spam accounts. Our findings indicate that 28.7% to 43.9% of these spam accounts remain active as of this writing. Furthermore, from the identified scams, we extracted 327 URLs associated with giveaway scams, 808 malicious contracts in arbitrage bot scams, and 429 scam cryptocurrency addresses. By analyzing the transaction history of the extracted scam addresses, we estimated that over 9,717 victims fell prey to these scams, resulting in a loss of up to 3.8 million USD. Overall, this study sheds light on the tactics, scale, and impact of cryptocurrency scams on social media and blockchain platforms, emphasizing the urgent need for effective detection and prevention mechanisms to protect users from these fraudulent activities.