Scientists used AI to find and track 95K ‘cryptocurrency free giveaway’ scams on Twitter

Researchers from San Diego State College in California developed a synthetic intelligence (AI) system to establish, observe and expose free cryptocurrency giveaway scams on X (previously Twitter). 

Referred to as GiveawayScamHunter, the automated system found 95,111 rip-off lists between June 2022 and June 2023 that have been created from 87,617 accounts on the X social community.

The researchers used the software to autonomously extract web site and pockets addresses related to the scams. In doing so, they have been in a position to accumulate 327 rip-off giveaway web domains and 121 new scam-related cryptocurrency pockets addresses.

Associated: Blockchain Capital’s X account hacked to promote token claim scam

Step one to approaching the issue concerned figuring out a brand new vector of assault for cryptocurrency giveaway scams: Twitter Lists. As a result of permissionless nature of the Lists characteristic on the social community, it presents a easy networking software for scammers to take advantage of.

To find out which lists handled giveaway scams, the group skilled a pure language processing software on knowledge from beforehand recognized giveaway scams.

The researchers have been in a position to establish almost 100,000 cases of giveaway rip-off lists utilizing this methodology, which allowed them to compile knowledge on beforehand unreported rip-off web sites and wallets.

Utilizing this knowledge, the group gleaned quite a few insights into how these scams unfold, how scammers goal victims and the approximate variety of victims scammed throughout the one-year examine interval.

Per the paper:

“By monitoring the transactions of the rip-off cryptocurrency addresses, this work uncovers that over 365 victims have been attacked by the rip-off, leading to an estimated monetary lack of 872K USD.”

The scientists reported their outcomes and the related accounts, domains and pockets addresses to each X and the cryptocurrency/blockchain group. Nonetheless, based on their paper, 43.9% of the related accounts stay energetic as of its Aug. 10 publication — although the researchers do observe that almost all of those are seemingly spam accounts not in energetic use.