<p>Artificial Intelligence (AI) has shown a significant presence in different day-to-day tasks. Natural Language Generation (NLG) is one of the areas where the powerful impact of AI can be clearly observed. The high efficiency of Large Language Models (LLMs) in producing texts similar to human-created text poses various challenges like fake reviews, fake news, and unethical practices in academics etc. Thus, it provides researchers with a new dimension to detect AI-generated text. In academics, some software, for example, Turnitin, provide information about AI-generated text, but they function like black boxes and not available for common use. Thus, there is a need to explore intensive pattern analysis to find the parameters on the basis of which AI-generated text can be detected. In this study, we have done a comparative examination of real and AI-generated sentiment data. The financial data, publicly available on Kaggle, is used for our study. AI-generated data is generated with the help of ChatGPT (GPT4 Model) using the existing financial data. Different linguistic quality parameters, sentiment-analysis performance parameters, complexity and diversity measures, and context similarity are evaluated for both datasets. The purpose of the exploration work is to recognize parameters that can differentiate real data from AI-generated data. For this intention, we evaluate both datasets on distinct methodologies and comparison dimensions. We identified 35 parameters used for linguistic analysis in the existing literature. Therefore, we evaluated 35 parameters for the comparison of real and AI-generated data and found 16 relevant parameters. These parameters will help in developing a model for detecting AI-generated text in future work.</p>