Test Case of a Chatbot-Based Virtual Assistant to Identify and Classify Cases of Harassment Within the University and Refer to the Student Welfare Department Using GPT-3

Harassment is a pervasive issue in university environments, with alarming rates reported globally. This paper presents a pioneering approach using GPT-3, a natural language model, to develop a chatbot-based virtual assistant aimed at identifying and categorizing harassment cases within universities. Extensive research highlights the prevalence of harassment, particularly affecting female students. The proposed methodology encompasses information collection, architecture design, assistant testing, and results evaluation. The virtual assistant demonstrates remarkable effectiveness in delivering relevant and compassionate responses to user messages, offering empathy, personalization, and the ability to detect violent language. While this innovation shows promise in addressing harassment issues, human validation, ethical considerations, and transparency remain paramount to ensure user safety and privacy. This technology, when judiciously employed in conjunction with human intervention, has the potential to enhance the emotional well-being of university students, exemplifying the fusion of technology and emotional support within an academic setting.