Publication

Abstract : Human trafficking remains a pervasive global issue, affecting millions of individuals through various forms of exploitation and coercion. The increasing use of digital platforms by traffickers to recruit and exploit victims necessitates the development of innovative and effective solutions. This research proposes a novel approach utilizing machine learning (ML) to identify suspicious activities indicative of human trafficking on social media platforms and within online job advertisements. This research proposes a novel approach to identify suspicious job advertisements that may indicate human trafficking. By analyzing keywords and their semantic relationships, we develop a system to flag high-risk postings. A multi-lingual dataset, comprising texts in English, Hindi, and Bengali, was created and meticulously labeled to determine the presence of suspicious content. This dataset was subsequently used to train a machine learning model capable of detecting potential trafficking activities in real time. The ultimate objective is to integrate this model into a web platform or browser extension, facilitating continuous monitoring and providing alerts to users. Moreover, the study introduces a robust model designed to detect patterns in job advertisements that are indicative of potential trafficking risks. By leveraging advanced natural language processing (NLP) techniques, specifically BERT embeddings, in combination with red flag detection mechanisms and sentiment analysis, this research presents a comprehensive framework for identifying and mitigating suspicious job postings. The findings underscore the critical role of leveraging advanced machine learning techniques to safeguard vulnerable populations, offering a scalable and automated solution to combat the evolving threat of human trafficking in the digital age.