This umbrella review aims to map the teaching methods used to integrate Artificial intelligence (AI) Literacy into the K-12 system and to highlight the gaps in the literature on the subject. Scopus, Eric and Web on Science datasets were used for the search and a total of 12 articles published between 2022 and 2024 were reviewed. The results show that project/problem-based learning is the most widely used teaching approach for AI Literacy courses in Secondary School, while robotics and games are more popular in the lower school grades. In terms of learning content, the majority of curricula focus on basic AI knowledge while ethics and social impact are mentioned less frequently. Moreover, hardware (e.g. robots, sensors) and software (e.g. Scratch, Arduino) are the most commonly used learning tools.

In particular, the umbrella review reveals the need to set up training courses for teachers, who often feel insecure about AI. Finally, the results indicate how important is promoting research especially into assessment criteria and methods, and to continue the development of interdisciplinary AI Literacy programmes.

Keywords: Artificial Intelligence, AI Literacy, AI Education, Teaching practice, Twenty-first century skills, K-12 Education.