Artificial Intelligence is widely used across contexts and for different purposes, including the field of education. However, a review of the literature showcases that while there exist various review studies on the use of AI in education, missing remains a review focusing on science education. To address this gap, we carried out a systematic literature review between 2010 and 2021, driven by three questions: a) What types of AI applications are used in school science? b) For what teaching content are AI applications in school science used? and, c) What is the impact of AI applications on teaching and learning of school science? The studies reviewed (n = 22) included nine different types of AI applications: automated assessment, automated feedback, learning analytics, adaptive learning systems, intelligent tutoring systems, multilabel text classification, chatbot, expert systems, and mind wandering detection. The majority of the AI applications are used in geoscience or physics and AI applications are used to support either knowledge construction or skills development. In terms of the impact of AI applications, this is found across the following: learning achievement, argumentation skills, learning experience, and teaching. Missing remains an examination of learners' and teachers' experiences with the use of AI in school science, interdisciplinary approaches to AI implementation, as well as an examination of issues related to ethics and biases.