Artificial Intelligence Neutrality: Framing Analysis of GPT Powered-Bing Chat and Google Bard

Abstract

This study examined neutrality in Artificial Intellignce (AI)-based tools in information dissemination, specifically Microsoft's GPT powered Bing Chat and Google's Bard, across three geopolitical topics. The Gamson and Modigliani framing model was employed for evaluation of the AI responses. Results revealed that while efforts were made to maintain neutrality, subtle Western, specifically American, perspectives persisted in the narratives, challenging the prevalent perception of AI as a neutral technology. These findings indicate that geographical and cultural contexts of AI development may inadvertently influence narrative framing. The study emphasizes the need for incorporating diverse perspectives in AI development and further research into this sphere. Limitations include the interpretive nature of the framing model and potential influence of cultural and academic backgrounds.