Truth Lens: An AI-Driven NLP Framework for Mental Stress Detection with Multilingual and Language Diversity Analysis

Mental stress has emerged as a significant concern among students and working professionals, driven by academic demands, occupational pressure, and modern lifestyle factors. Timely detection of stress is crucial for preventing long-term mental health complications. Recent advances in Artificial Intelligence (AI) and Natural Language Processing (NLP) have enabled automated analysis of textual data to identify psychological states through linguistic patterns. This paper presents Truth Lens, an AI-driven NLP framework for mental stress detection with an emphasis on multilingual processing and language diversity. The study systematically examines existing transformer-based models and benchmark datasets to evaluate their effectiveness in stress classification tasks. Multilingual transformer architectures, particularly mBERT-based models, are analyzed due to their capability to capture cross-lingual semantic representations. Findings reported in prior experimental studies indicate that such models achieve promising performance, with accuracy and F1-score values of approximately 73% and 77%, respectively, on real-world social media text data. The results underline the potential of multilingual NLP approaches for scalable and inclusive mental stress detection. This work provides a conceptual foundation for extending AI-based stress analysis systems to low-resource and regional languages, supporting the development of accessible mental health monitoring solutions.