Abstract:

Generative artificial intelligence (AI) employs sophisticated algorithms, such as deep learning and reinforcement learning, to generate diverse content, including text, images, and audio, requiring minimal human input. Its capacity to produce authentic and realistic outputs makes it valuable across sectors, like healthcare, finance, and entertainment. In the scientific field, GAI enhances productivity by supporting scientists in idea and hypothesis generation, data pattern analysis, experiment design, and research paper writing, despite concerns over AI-generated information's accuracy and issues of authorship and ownership.

As AI permeates modern societies, its impact is expanding to education. It is promising personalized learning experiences that cater to individual student needs and preferences, thereby being presumed to enhance educational outcomes. The impact of generative AI on education is claimed to include numerous opportunities. For example, in computer education, natural language processing (NLP) can personalize learning experiences, aid in student recruitment, and provide solutions for programming exercises. However, its implementation in education also presents unique challenges. For example, it risks being viewed as the “ultimate epistemic authority”, where a single truth is assumed without enough evidence or qualifications.

This work explores the opportunities for incorporating generative AI into science teacher education. Thus, it focuses on one type of public that may be understudied in this field. The research question is: What are the potential opportunities of incorporating generative AI into science teacher education presented in the literature? To answer this question, the Preferred Reporting Items for Systematic Reviews (PRISMA) statement was used. The search terms "generative artificial intelligence" OR "artificial intelligence" AND "science teacher education" OR "science teacher training" OR "teacher education" OR "teacher training" AND "opportunities" OR "benefits" OR "advantages" were used in the Scopus database, in March 2024. This search retrieved 34 registers, published between 1987 and 2023.

This review includes peer-reviewed published empirical studies, published as an article, conference paper, or book chapter, and focused on the potential exploration of (generative) AI in (science) teacher education. Due to the novelty of the topic, studies related to non-generative AI and studies in teacher education (not specifically in science) were considered. Hence, after the application of the inclusion criteria, 18 registers were excluded from this study.

The selected studies were analyzed and synthesized to develop a coherent understanding of the opportunities AI presents to teacher education, such as teachers’ favorable attitudes toward AI education for teaching and for future use in their practices, or skills practice opportunities for teacher candidates.

Despite the promising outlook, this study reveals a gap in the literature concerning teacher training. Moreover, the scarcity of AI-learning opportunities in teacher initial education programs makes prioritizing the development of AI-supported education crucial. This empowers teachers to make informed decisions about incorporating AI into educational practices, revealing the need for further research to enhance understanding and guide best practices, potentially advancing science education and effective AI integration in science teaching.

Keywords:

Artificial intelligence, teacher education, systematic literature review.