This post is also available in:
עברית (Hebrew)
A recent study suggests that advanced technologies like ChatGPT could enhance anti-terrorism efforts by profiling potential terrorists and assessing their likelihood of engaging in extremist activities. Published in the Journal of Language Aggression and Conflict, the study titled “A Cyberterrorist Behind the Keyboard” explores how automated text analysis can aid in threat assessment.
Researchers from Charles Darwin University (CDU) analyzed 20 public statements made by international terrorists post-9/11 using the Linguistic Inquiry and Word Count (LIWC) software. Then, they employed ChatGPT to examine selected statements from four terrorists, asking it to identify key themes and underlying grievances. The results showed that ChatGPT could effectively pinpoint central themes such as retaliation, opposition to democratic systems, and anti-Western sentiment.
The identified themes were strikingly comprehensive, including ideas of martyrdom, dehumanization of opponents, and grievances against multiculturalism and mass immigration. ChatGPT also highlighted motivations for violence, such as desires for retribution, religious grievances, and feelings of cultural oppression.
Crucially, these themes aligned with indicators from the Terrorist Radicalization Assessment Protocol-18 (TRAP-18), a framework used by authorities to evaluate individuals at risk of engaging in terrorism. According to TechXplore, Lead author Dr. Awni Etaywe, an expert in forensic linguistics, emphasized the complementary role of large language models (LLMs) like ChatGPT: “While LLMs cannot replace human judgment or close-text analysis, they offer valuable investigative clues, accelerating suspicion and enhancing our understanding of the motivations behind terrorist discourse”.
This study highlights the promise of integrating machine learning into proactive forensic profiling. Co-authored by CDU’s Dr. Kate Macfarlane and Professor Mamoun Alazab, the paper calls for further research to refine the accuracy and reliability of LLMs in threat identification.
In this day and age, where terrorism is a growing phenomenon in Western countries, there is significant importance in continuing to develop this area to better identify potential threats.
