Generative artificial intelligence (GenAI) tools, such as ChatGPT and Med-PaLM, are revolutionizing how individuals access health information and conduct self-diagnosis. Unlike traditional symptom screening tools, these tools generate context-dependent, personalized conversational responses. As the use of these tools increases, research is now beginning to examine their advantages and disadvantages. Recent studies emphasize the desirable features of accessibility and greater user engagement, while also addressing issues of misinformation, anxiety, and procrastination in seeking clinical help. Existing studies tend to be fragmented, with little integration between technical, psychological, and ethical perspectives. This systematic review brings together recent evidence to examine the benefits and the impact of GenAI on the risk diagnostic accuracy and user behaviors. The research highlights key gaps, including how to implement plans in the real world, equitable access for all, and the requirement for regulations. It also demonstrates that GenAI has the potential to significantly enhance public health and patient care. Nevertheless, the research also considered fundamental questions regarding safety, accountability, and ethical harm. The current study discussion recognizes how important it is to undertake wide policies and research across different disciplines on the use of these technologies in medicine. These policies should be able to guarantee that GenAI is used efficiently, ethically, and address the varied needs of users.