Fighting fake news in the age of generative AI: Strategic insights from multi-stakeholder interactions

The advancements in algorithm technology have led to a proliferation of artificial intelligence-generated fake news, resulting in significant social harm. Promoting multi-stakeholder engagement in fake news governance is beneficial for establishing a robust information ecosystem. The primary stakeholders, including the government at the policy-making end, user-generated content platforms at the algorithm development end, and opinion leaders at the news dissemination end, possess varying degrees of initiative and roles in governance. The main objective of this study is to investigate the evolutionary process of behaviors among multi-stakeholders in fake news governance and their influencing factors under different news environments. This study constructs an evolutionary game model to identify the conditions for the realization of five models of fake news governance. Stakeholders' behaviors in different states are affected by external factors, such as news environment, penalties, and incentives, as well as internal factors, such as governance capability deficiencies and platform algorithm reliability. The research findings expand the boundary of adaptive governance theory by revealing the mechanisms of stakeholder collaboration and the interaction between stakeholders and the news environment in fake news governance. These insights offer valuable guidance for advancing the transformation and enhancement of fake news governance models.