Researchers from San Diego State University have developed an artificial intelligence (AI) system called GiveawayScamHunter to detect cryptocurrency giveaway scams on Twitter. The automated system identified a shocking 95,111 scam lists created by 87,617 accounts on the social media platform between June 2022 and June 2023.
Using GiveawayScamHunter, the researchers were able to extract website and wallet addresses associated with the scams, uncovering 327 scam giveaway internet domains and 121 new scam-related cryptocurrency wallet addresses.
The researchers provided valuable insights into the workings of these scams, including the targeting of victims and the estimated number of victims scammed during the one-year study period. The report estimated that over 365 victims were attacked by crypto scams during that period, resulting in a loss of more than $872,000.
Interestingly, the researchers found that Twitter Lists, a networking tool on the platform, have become a convenient vector for scammers to exploit due to their permissionless nature. To identify which lists were involved in giveaway scams, the team trained a natural language processing tool on data from previously identified scams. This allowed them to identify almost 100,000 instances of giveaway scam lists and collect data on previously unreported scam websites and wallets.
The researchers have shared their findings, along with the associated accounts, domains, and wallet addresses, with both Twitter and the crypto community. However, as of the publication of their paper on August 10, 43.9% of the associated accounts remained active.
The use of AI in detecting cryptocurrency giveaway scams on social media platforms is significant as scammers and bad actors have already started utilizing AI to develop new methods of fraud and deception. By leveraging AI-powered tools, scammers can amplify their reach and create a seemingly loyal fanbase.
These fake accounts and interactions give the illusion of credibility and popularity to their scam projects. Scammers may also employ AI-driven chatbots or virtual assistants to engage with individuals, provide investment advice, promote fake tokens and initial coin offerings, or offer high-yield investment opportunities. The use of AI can also challenge social proof-of-work, as it assumes that crypto projects with greater and more loyal followings online are legitimate.
One example of how scammers are using AI is through “pig butchering” scams, where AI instances spend several days befriending vulnerable individuals, only to scam them in the end.
