The fashion e-commerce industry faces significant challenges due to biases embedded in consumer reviews, which can distort product perceptions and negatively impact consumer trust. These biases, ranging from race and size to class and gender, affect both consumers’ experiences and businesses’ reputations. This study introduces a novel methodology for detecting and analyzing bias patterns within large-scale fashion e-commerce reviews using AI-driven sentiment analysis and network science techniques. We leverage large language models (LLMs) to extract bias-related terms from domain-specific literature and apply sentiment analysis to map these reviews onto a bipartite network. By visualizing the consumer review network, we identify key clusters where biases intersect, revealing insights into how consumer sentiments vary across different bias categories. Our findings highlight the complexities of intersectional biases, such as race, size, and class, and demonstrate how positive and negative reviews cluster around these biases, impacting consumer perceptions. Additionally, the results show that reviews tied to religious biases form tightly-knit communities, while size and class-related biases exhibit more widespread influence across the network.