Artificial Culture : An explorative study on OpenAI’s ChatGPT and its cultural intelligence

Abstract

[""Studies proving that ChatGPT is susceptible to bias shed light on the importance of creating scientific guidelines for culturally sensitive and inclusive AI-systems. While general cultural incongruencies in AI have already been identified, the question of AI’s ability to navigate the culture of its users remains unexplored territory. The topic of this thesis therefore consists of two components: (1) artificial intelligence and (2) cultural intelligence, coming together as a study exploring the extent and implications of ChatGPT’s cultural intelligence. The employed methodology is composed of three experiments in which the first two assess and reassess ChatGPT’s cultural intelligence (CQ) based on a questionnaire, while a third experiment tests ChatGPT’s live CQ when confronted with intercultural critical incidents. Incorporating primarily evidence from the aforementioned experiments, this study demonstrates that ChatGPT’s cultural intelligence is rather pronounced with a specific focus on certain subdimensions of CQ. ""]