The Kaplan-Meier estimator, commonly used in various fields, has been reinterpreted as an M-estimator by maximizing a specific function. This new perspective allows for a different way to understand and calculate the estimator. The study shows that the traditional Kaplan-Meier estimator and the new M-estimator are equivalent, with similar results when applied to real datasets. This new approach also allows for the establishment of confidence intervals and bands for the estimator.