The paper provides a systematic review of chatbot history and current technologies and their social effects and operational difficulties. The research combined findings from 45 peer-reviewed studies which used PRISMA guidelines to analyze multiple academic fields. Chatbots have progressed from basic rule-based systems like ELIZA and PARRY to advanced AI systems which use machine learning and natural language processing (NLP) and large language models (LLMs). The development of Retrieval-Augmented Generation (RAG) and multimodal interaction and emotion-aware models enables chatbots to generate human-like responses while maintaining context understanding. The expanded range of chatbot applications now includes healthcare and education and finance and government and commercial organizations which benefit from fast responses and enhanced worker efficiency. However, persistent issues remain in security vulnerabilities, ethical bias, multilingual accessibility, data privacy, and the absence of standardized evaluation frameworks. The research demonstrates that digital communities need ethical management systems and protected integration protocols and user-focused design principles to function properly. The development of future chatbots needs to prioritize emotional intelligence and inclusiveness and transparency to establish public trust for digital society expansion.
Casanova et al. (Thu,) studied this question.