Text summarization is the process of condensing lengthy texts into concise and coherent summaries, capturing the main points of the document. It presents a significant challenge in machine learning and natural language processing (NLP) due to the vast volume of digital data available. There is a growing demand for algorithms capable of automatically condensing extensive texts into accurate and understandable summaries to effectively convey the intended message. Machine learning models are typically trained to comprehend documents, extracting essential information to produce the desired summarized output. The application of text summarization offers several advantages, including reduced reading time, accelerated information retrieval, and efficient storage of more information. In NLP, two primary methods for text summarization exist: extractive and abstractive. The extractive approach identifies key phrases within the source document and assembles them to form a summary without altering the text’s content. The abstractive technique paraphrases and condenses sections of the source document. In deep learning applications, abstractive summarization can overcome grammar inconsistency issues often encountered in the extractive method. Our analysis has examined various existing methods for text summarization, including unsupervised, supervised, semantic, and structure-based approaches, and has critically assessed their potential and limitations. Specific challenges highlighted include anaphora and cataphora problems, interpretability issues, and readability concerns for long texts. To address these challenges, we propose solutions aimed at improving the quality of the dataset by addressing outliers through the integration of corrected values obtained from human-generated inputs. As research in this domain progresses, we anticipate the emergence of innovative breakthroughs that will contribute to the seamless and accurate summarization of lengthy textual documents.