This paper assesses the extent of political polarization in the United States by demonstrating the phenomenon of political identity projection, where individuals attribute political affiliations to others based on political discourse. This aspect of political behavior, often found in interactions between authors on various social media platforms, remains relatively unexplored. To address this gap, our research utilizes a comprehensive dataset of comments on YouTube news videos from three prominent US cable news networks (Fox News, CNN, and MSNBC) to interpret expressions of political polarization. First, we assess the accuracy of LLMs in identifying political identity projections, exploring the potential biases these models may incorporate. Second, we conduct a user engagement analysis that highlights interaction patterns and their implications for understanding political identity projections across different news outlets.