Detection of Fraudulent Internship Opportunities Using Machine Learning Techniques

Online internships have become a common pathway for students to gain industry exposure and practical skills. Along with this growth, fraudulent internship offers have also increased, often misleading students through unrealistic benefits, misleading descriptions, or fake certifications. Due to the large number of internship listings available online, manually verifying such listings for fraud is time-consuming and unreliable. This research presents a machine learning-based approach to classify internship postings as legitimate or fraudulent using textual information. A self-developed dataset containing real and fake internship descriptions was created and processed using standard natural language processing techniques. Textual features were extracted using the TF-IDF method, and a Logistic Regression model was trained for classification. The system was evaluated using accuracy, precision, recall, and F1-score metrics. Experimental findings demonstrate that the proposed approach can successfully identify fraudulent internship postings, indicating that machine learning can serve as an effective tool for reducing internship-related scams and protecting students.