AI chatbots powered by large language models (LLMs) are widely used as recommender systems. While accelerating access to information, AI can also give factually incorrect responses or propagate societal biases.
Who does AI consider an expert? New benchmark shows bias across 22 LLMs
