This study addresses the fundamental challenge of Large Language Model (LLM) identification using a comprehensive text classification system and Responsible Artificial Intelligence (RAI) Principles. The main objective is to create a reliable, transparent, and ethically responsible model that can differentiate between text created by AI and those written by humans according to ethical AI deployment standards. For the classification problem, we used deep learning models such as GPT, BERT, and LSTM with richer text representation achieved by using word embedding methods such as Glove and FastText. The proposed methods encompass multiple approaches to ensure the model’s accuracy and ethical integrity. Explainability is achieved through SHAP and LIME for transparent interpretation of predictions. Robustness and reliability are enhanced through adversarial testing and uncertainty estimation via Monte Carlo simulations. Cross-validation techniques ensure consistent model performance while ensemble methods further improve prediction accuracy and generalization. Privacy, security, and fairness considerations are incorporated to align with ethical standards and mitigate potential biases. The model achieves 99.58% of accuracy, 99.30% of precision, 99.55% of recall, and an AUC of 99.98% on a public LLM-detection dataset available on Kaggle. These results show how well this model recognizes LLM-generated content