The article I’ve chosen to review is “Digital Fingerprinting for Identifying Malicious Collusive Groups on Twitter” by R. Ikwu, L. Giommoni, A. Javed, P. Burnap, and M. Williams. When looking through this article, I noticed that it related to the social science principles in a few ways. One of which was relativism, because when reading I noticed that it talks about how Twitter created a way to make it easier for malicious actor to hide scam links within twitter URLs as well as malicious actors using bot accounts to spread misinformation as well as spreading the scam links. The article also uses Parsimony to explain how the data was collected to characterize a malicious actor’s digital footprint by checking the URL fingerprint to check how often certain tools are used within the URL such as the host IP as well as the account fingerprint to find out the account’s creation time, location, or age. The also related objectivity to it by using certain hashtags relating to COVID-19 and finding out which ones had links that could be potentially malicious.
The questions that come up with this article is “How can we tell which shortened URLs are malicious?” and “How can we tell if an account is malicious?” because the article explains how they’re able to find out which accounts are malicious via links and by the stats on the accounts. The hypothesis I would say they had created would be that the malicious accounts are found based on the age of the account as well as what their tools give based on the links they found on certain tweets. They used a method of research that included using hashtags during Covid’s timeframe in 2020 to 2021. Another research method they used was they used a program called VirusTotal to run Twitter’s shortened URLs through over 80 different antivirus programs to scan if a link is malicious or not by labeling them if the scan claims as malicious. The third method they used is to create the digital fingerprints for the malicious actors via URL, account data, language, and activity type. One of the data types that they found and created was a set of tables made of URL fingerprint features, tweet account characteristics, text language characteristics, and account activity. Another type of data they found was a table that determined if URLs were malicious or benign via their digit count, path length, and number of capitalizations. The analysis they made was by using data cleaning to remove tweets with unneeded info, one-bot encoding to creating categories, feature scaling, and feature selection via a graph explaining clusters found within data fingerprints.
The article relates to archival research because with how they’re performing their research, the create an archive of twitter posts with specific search criteria and it has a potential to deal with ethical issues although the information is easily accessible. It also relates to multi-method research because they’re collecting the data through a field study by using the antivirus software while also performing hands on searches with Twitter’s search function to avoid finding unethical searches. It also relates in an interdisciplinary matter because it could deal with diversity due to those who make the accounts and what part of the world they work within as well as their social status. The article also relates to psychology because there’s a portion in the background section of the article that explains how people on Twitter interact with one another to gain information about how certain accounts can become more friendly with others, leading to trust and scamming becoming easier.
This topic concerns those who were affected by Covid because those who were stuck inside that are unaware of the problems that come with altered Twitter links could be in danger if they were to click the links. This also affects those who don’t know much English because someone could inform them that it’s a malicious link and they wouldn’t know, so they click it anyway. It also concerns those who have no idea of the trouble they’re causing by joining these malicious groups due to supposed social engineering or having their accounts taken over by falling for scams via social engineering.
This study contributes to helping to inform those who are unaware of scam links on twitter to make sure they check a link with their antivirus before they click on it because if they click on it they could potentially lose their account or lose access to their entire library of accounts and money. It also contributes to informing Twitter about how many bot accounts or scam accounts lurk on the site that can potentially harm those on the outside world.
Works Cited
Ruth Ikwu, Luca Giommoni, Amir Javed, Pete Burnap, Matthew Williams, Digital fingerprinting for identifying malicious collusive groups on Twitter, Journal of Cybersecurity, Volume 9, Issue 1, 2023, tyad014, https://doi.org/10.1093/cybsec/tyad014