Skip to main content

L’analyse des traces langagières comme vecteur de détection de relations entre des cas de fraudes en ligne

Clara Degeneve ; Julien Longhi ; Quentin Rossy (2026) — Criminologie, Forensique et Sécurité4(1):8332

Citation activity

Total citations · Google Scholar
Unavailable
Average per calendar year
Citation count unavailable

Counts reflect Google Scholar’s coverage and do not measure research quality. Source, calculation and limitations

View Original Source (opens in a new tab)

DOI: 10.26034/la.cfs.2026.8332 (opens in a new tab)

Country:   Global
Tags:   Crime Analysis, Fraud Scripts, Linguistic Traces, Online Convergence Spaces, Repeated Segment Analysis, Romance Scams, Scambaiting, Template Phrases
What do these research terms mean?
Preprint
A manuscript shared before formal peer review and publication. Check whether a later published version is available.
Dataset
A collection of data or examples for others to inspect or reuse. It can appear in Library search and topic mapping, but RSRC does not use dataset records as evidence in Research Insights.
Dissertation or thesis
Research submitted for an academic degree. This describes its format, not its reliability.
Journal article
An article published in a journal. This label alone does not establish peer review, study quality, or how well the findings apply elsewhere.
Qualitative research
Examines experiences, meanings, or processes, often through interviews or observations. It can explain how something happens without estimating how common it is.
Quantitative research
Uses numerical measurements to describe patterns or test relationships. A relationship between two measurements does not by itself show that one causes the other.
Systematic review
Uses a planned, documented method to find and assess research addressing a question. Its conclusions still depend on the included studies and what the search covered.
Meta-analysis
Statistically combines results from multiple studies. Combining studies does not remove weaknesses in their design or make unlike populations interchangeable.
Not classified
This record has no recognized label in this filter. It does not mean the publication used no method, or that no research exists.

Definitions draw on DataCite resource types; Cochrane review methods; NLM: association and causation. RSRC’s dataset and classification rules are explained in our methodology.

Citation tools
Download RIS Download BibTeX 


              
              

            

          

        


      

      

    

        
Transparency

        
Evidence and review status

        
This page contains AI-generated content. No human content review or subject-matter-expert review is recorded.

    


    

                    

                
Source basis

                

                    Downloaded PDF
                

            

                    

                
Source updates

                

                    No notice found at last check
                

            

                    

                
AI-generated page content

                

                    Yes
                

            

                    

                
Automated checks

                

                    Passed
                

            

                    

                
Administrative approval

                

                    Yes
                

            

                    

                
Human content review

                

                    Not recorded
                

            

                    

                
Subject-matter-expert review

                

                    Not recorded
                

            

            


    

        
            How this was prepared
        
        

                            

                    
Source basis

                    

                    
RSRC downloaded and privately stored a copy of the paper for internal analysis. The PDF is not offered to viewers from this page.


                                            
    
                                                            
  • 
                                    PDF added to RSRC:
                                    
                                
    • 
                                                    

                    
                                        

                

                            

                    
Source updates

                    

                    
No incoming update notice was found in the dated Crossref response. Coverage is incomplete, particularly for corrections and expressions of concern; this is not a guarantee that the source is valid or unchanged.


                                            
    
                                                            
  • 
                                    Last source-status attempt:
                                    
                                
    • 
                                                    

                    
                                        

                

                            

                    
AI-generated page content

                    

                    
AI-generated research notes displayed on this page: Synopsis, Identified gaps, Methods, Limitations, Future work. The paper itself is not described as AI-generated.


                                            
    
                                                            
  • 
                                    Document analysis recorded:
                                    
                                
    • 
                                                            
  • 
                                    Synopsis generation recorded:
                                    
                                
    • 
                                                            
  • 
                                    Page record updated:
                                    
                                
    • 
                                                    

                    
                                        

                

                            

                    
Automated checks

                    

                    
The current, source-bound synopsis passed the recorded versioned publication checks.


                                            
    
                                                            
  • 
                                    Checks completed:
                                    
                                
    • 
                                                    

                    
                                            

                            
                                View passed checks (2)
                            
                            
    
                                                                    
  • Length, completeness, repetition, refusal, boilerplate, and active-markup screening
    • 
                                                                    
  • Numerical claims checked against the available source text
    • 
                                                            

                        

                                        

                

                            

                    
Administrative approval

                    

                    
An authenticated administrator approved the bibliographic record for public Library display. This is not a review of every research claim.


                                            
    
                                                            
  • 
                                    Approved for public display:
                                    
                                
    • 
                                                    

                    
                                        

                

                            

                    
Human content review

                    

                    
No human review is recorded for the AI-generated content displayed on this page.


                    
                                        

                

                            

                    
Subject-matter-expert review

                    

                    
RSRC has not recorded review of this content by a subject-matter or methods expert.


                    
                                        

                

                    

    


    

        Review-state definitions

        Found a possible error?
        Request a correction.
    



    



              

        
                  

            
Synopsis

            
This publication investigates whether linguistic traces in online fraud communications, specifically romance scams, can reveal connections between separate fraud cases. The authors frame language traces as textual elements within illegal or disputed texts that may illuminate not only who is involved but also how the fraudulent activity operates within a broader system. The study aims to detect relationships among romance-scam cases by examining language segments across conversations, rather than classifying individuals as legitimate or fraudulent. Using a repeated-segment analysis on a corpus of online conversations, the researchers identify template phrases and recurring script structures that recur across multiple cases. They find that a large portion of detected links rely on content tied to everyday life or to non-model phrases, while a subset of longer, more distinctive phrases points to shared practices among fraudsters. The analysis also uncovers online platforms where these scripts are shared, suggesting virtual spaces that facilitate diffusion and adaptation of fraudulent templates. In total, the study reports hundreds of links among dozens of cases, with most connections arising from non-model language and a smaller set from more distinctive dialogue. The authors acknowledge important caveats: many repeated segments are not uniquely identifying and can be highly generic, limiting inferential strength when linking cases. They caution against over-interpreting associations, particularly given evolving scripts and the potential influence of generative AI on language. They also note that the dataset reflects a snapshot from 2021 and that broader applicability requires ongoing updating and contextualization. Overall, the work proposes linguistic-trace analysis as a promising, but prudently applied, method to illuminate the ecosystem and interconnections underlying online romance fraud.

          

        
                  

            
Identified Gaps

            
The paper identifies no prior linguistic study focused on detecting relationships between romance-fraud cases through language traces. It also notes limited access to data directly produced by fraudsters, since available corpora commonly rely on victim accounts or police reports. Existing work emphasizes persuasion, victimization, or fraudulent-profile detection rather than links among cases and offender knowledge-sharing systems.

          

        
                  

            
Methods

            
The study analyzed 180 scambaiting conversations with suspected romance fraudsters associated with simulated victims located in Switzerland, Canada, and France. Researcher messages and system messages were removed; identifying information was manually anonymized. The authors compared identifying information, then used Python Textacy to detect repeated message segments across cases. ChatGPT-4o provisionally categorized segments, followed by manual review and filtering. Characteristic repeated segments were searched on Google to identify online template-sharing platforms.

          

        
                  

            
Limitations

            
The scambaiting corpus reflects interactions shaped by researchers seeking information and does not necessarily represent genuine romance-seeking victims. Conversations did not extend beyond money requests, limiting coverage of later fraud stages. Repeated segments were often generic, and isolated text can weakly support inferences about direct case relationships. ChatGPT classification required substantial manual correction and remains partly subjective. Shared templates across highly visible sites prevent attribution to one specific platform or offender community.

          

        
                  

            
Future Work

            
Reassess language-model classification using the preceding and following messages, and use multiple independent human reviewers. Regularly update repeated-segment databases and develop a chronology of circulating templates. As generative AI may replace static scripts, investigate AI-style detection, nonhuman conversational patterns, and computational signatures associated with models used by fraudsters.

          

              

    
    

        
Related Research Insights

        

            See how this publication connects to RSRC's living evidence syntheses through current citations and research-topic mapping.