The social media content often contains inherent biases, posing significant challenges in the development of fair machine learning systems. These biases negatively impact machine learning models, often leading to unfair and at times, discriminatory outcomes. In this paper, we present a novel method for bias detection in social media content using sentiment analysis and feature extraction techniques to identify potential sources of bias without relying on deep learning models. We test our method on several standard machine learning classifiers, such as Gaussian Naive Bayes, Decision Trees, and Support Vector Machines, using cited bias-annotated datasets with an accuracy of 85.64%.