Bias in AI: Analyzing Think Tank Ratings

This chapter examines a study on how 26 influential U.S. think tanks are rated by leading large language models, revealing a significant ideological bias favoring center-left organizations. It discusses the implications of this bias on research integrity and public engagement, highlighting potential feedback loops that could reinforce political disparities. The chapter also encourages model builders to conduct bias audits to promote more balanced representations in AI outputs.