Two Illinois Institute of Technology graduate students have published research examining whether extremists can be identified through their anonymous online posts using machine learning and open-source intelligence software.
Andreas Vassilakos (ITM/M.A.S. CYF ’21) and Jose Luis Castanon Remy (M.A.S. ITM 2nd Year) published “Illicit Activities Beneath the Surface Web: Investigating Domestic Extremism on Anonymous Social Media Platforms” in HOLISTICA Journal of Business and Public Administration. Dr Maurice Dawson, Illinois Tech assistant professor of information technology and management, and Tenace Kwaku Setor, assistant professor of information science and technology at the University of Nebraska Omaha, co-authored the paper.
The researchers examined online platforms such as Reddit and 4chan, where anonymous extremist rants and thoughts can be found easily. Domestic terrorists in California and New Zealand posted manifestos on these platforms before carrying out mass shootings. In each of these two cases, the shooters identified themselves as white nationalists and used these social media platforms to anonymously post their radical ideas and perceived viewpoints of population groups that conform to their own fanatic identities in political, ethnic, and social status.
“We collected actual messages from forums like Reddit and 4chan,” Vassilakos says. “Specifically, we reviewed subreddits [topic-based posts] that were focused on politically incorrect and racial context. Through these platforms, we were able to analyze data that was posted in plain text. We did not interpret the content, but collected it verbatim.”
The researchers used Open-Source Intelligence (OSINT) software, widely used by the United States government, to collect input values and data from the social media posts, which were then moved into a spreadsheet for analysis. By combining OSINT with artificial intelligence and machine learning techniques, the researchers hope to be able to identify anonymous posters.
“With this intelligence-gathering strategy, we can collect publicly available data to conduct our analysis,” Vassilakos says. “People are often not careful when they share data on the internet. Combining OSINT and other machine learning tools, we can excavate much information that can lead to valuable conclusions.”
After identifying these posts, an investigation into who originated the post can begin. Using tools such as Maltego, the researchers can examine IP addresses, MAC addresses, and mobile devices to unveil the identity of the poster.
Read more at Illinois Institute of Technology
