AI Literacy in Education across Regulatory Frameworks: A Comparative Analysis of the European Union, United States and China

AI literacy is no longer just an educational objective or an individual competence. State authority now uses it as a governance framework to manage risks and assign accountability for AI use in education. Yet, current literature continues to frame the concept almost entirely around user competence and ethical boundaries. This leaves its growing institutional role, specifically how it acts as a mechanism for legal responsibility, largely unexamined. To address this gap, the article examines how AI literacy is defined, framed, and translated into educational responsibility across three important, yet different governance models: the European Union, the United States and China. These frameworks were selected because they represent three different approaches to AI literacy governance in education: a legal-compliance model in the European Union, a decentralised, and incentive-based model in the United States, and a centralised, state led model in China. Through a qualitative comparative analysis of legal texts, official guidelines, and educationspecific policy documents, the study analyses how the same concept is assigned different regulatory and educational functions across three jurisdictions. The findings prove that AI literacy extends far beyond personal skills and competences, and it now operates primarily as a mechanism of institutional governance, fundamentally reshaping educational policy, teacher liability, and algorithmic responsibility.