A look inside the development of intelligence gathering framework for analyzing cyber extremism on social medial.

Introduction

This is a review of an article titled “Development of an Intelligence Gathering Framework for Analyzing Cyber Extremism on Social Media Networks.” It is a scholarly article published in Volume 18, Issue 1 of the International Journal of Cyber Criminology, covering January to June 2024. The premise of the article is to use intelligence gathering techniques such as Python web crawlers to collect data on social media platforms to gain and analyze data to attempt to identify and combat cyber extremism (Montasari et al., 2024). This review will assess the article through three concepts discussed in class: empiricism, objectivity, and skepticism.

Data

The article uses empirical data collection methods such as Selenium and Tweepy to conduct field research by scraping through posts from X. They are looking for keywords such as “attack”, “jihad”, “terrorism”, and the hashtag “#extremism” in addition to others that may be associated with extremists’ behavior (Montasari et al., 2024). This data was then stored in MongoDB, which they selected for its scalability and capacity to manage large amounts of unstructured social media data (Montasari et al., 2024).

Social Sciences

This article relates to social sciences in several ways. The research used field data for up-to-date empirical information. They empirically analyzed the data to see correlations between posts and keywords such as “kill” (appearing 991 times) and “gun” (Montasari et al., 2024). They used the data to attempt to map radicalization and recruitment patterns using targeted keyword associations (Montasari et al., 2024). Objectivity is present as the researchers maintain a neutral stance and focus on the data and facts. Skepticism appears in questioning keyword intent (e.g., whether “kill” always signals extremism) and opens the study to scrutiny of its keyword-based approach (Montasari et al., 2024).

Technology

The data was collected using several different types of technology including Python-based tools like Selenium and Tweepy. These tools scrape posts on social media with extremism-related keywords like “attack” and “jihad”, storing them in MongoDB, which was chosen for scalability and unstructured data handling (Montasari et al., 2024). Objectivity plays a part in this standardized method of data collection and storage. They used machine learning models like BERT (achieving 91% accuracy) and NLP for keyword analysis, which also enhances objectivity by using an unbiased software tool instead of human researchers. They still maintained their skepticism by having five experts annotate the data (achieving a Cohen’s Kappa of 0.82) (Montasari et al., 2024).

Conclusion

The article through its research and analysis methods does employ class concepts well. Empiricism is evident with real-world field data from X, objectivity through technology to remove human bias, and skepticism in using five experts to validate tools and data reliability. The article doesn’t really focus on marginalized groups, and using keywords to flag X posts doesn’t account for people who use them as a joke, or to describe a situation, or maybe just as hate speech. These types of studies are great for looking inside our society to try and predict behavior, but a worry arises when considering whether predicting an act of violence justifies intervention without due process. Where do we stop as a society, realizing we can’t punish people for potential actions without evidence? Rounding people up for online speech is a slippery slope.

References

Montasari, R., Carpenter, V., & Hill, R. (2024). Development of an intelligence gathering
framework for analysing cyber extremism on social media networks. International Journal of Cyber Criminology, 18(1), 1-25.
https://cybercrimejournal.com/menuscript/index.php/cybercrimejournal/article/view/387/108