Recent research has found that large language models consistently capture and replicate undesirable societal biases relating to race, religion, and gender. However, political bias is not well explored. This study investigates the political bias present in the state-of-the-art large language model GPT-3. To investigate political bias, I apply Natural Language Processing techniques to develop a political sentiment analysis model. Using this model, I analyze the ideological bias present in political essays written by GPT-3, finding that GPT-3 has a moderate left-leaning bias and tends to replicate the ideological bias of prompt text.



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.