Growing literature has shown that powerful NLP systems may encode social biases; however, the political bias of summarization models remains relatively unknown. In this work, we use an →