Exploring the Merchera Ethnic Group Through ChatGPT: The Risks of Epistemic Exclusion

Abstract

The study of underrepresented ethnic groups in the social sciences is often hindered by structural and epistemic barriers that limit access to culturally embedded knowledge. This article examines the potential of the GPT-4 version of the ChatGPT language model as a complementary research tool used to generate insights into the Merchera ethnic group, whose presence in the academic literature remains minimal and often characterised by misrepresentation. Through a comparative analysis considering ChatGPT responses and the scarce number of existing sources, this study explores the model’s reliability, depth, and limitations. The findings reveal that while ChatGPT offers a coherent synthesis of available knowledge, it tends to reproduce the prevailing biases and informational gaps found in the existing academic discourse. The paper concludes that generative AI may serve as a provisional support for research on marginalised communities, but its outputs must be interpreted with caution and situated within a framework of critical inquiry and ethical responsibility.