Researchers are leveraging the known issue of biases in artificial intelligence to gain insights into historical social attitudes. By training AI models on novels from specific decades, they can highlight the cultural prejudices prevalent during those times. This innovative approach reveals how biases have evolved and provides a unique perspective on the changing landscape of societal norms.
Large language models (LLMs) learn by analyzing extensive text datasets, often inheriting the biases contained within. This methodology not only sheds light on the prejudices ingrained in literature but also serves as a tool for understanding the evolution of cultural attitudes throughout history.
