Evolution of artificial intelligence in education from technology adoption to ethical governance and educator preparedness

As Artificial Intelligence (AI) reshapes teaching, learning, and administration, educators face pedagogical, ethical, and professional challenges. This study maps the structure, themes, and research trends of AI-in-education scholarship using bibliometric and quantitative content analysis. Using the PRISMA protocol for search and screening, 208 peer-reviewed articles from the Scopus database, published between 2021 and 2025, were retrieved, and analysed using Biblioshiny (Bibliometrix 5.5.0; R 4.6.1) and VOSviewer 1.6.21. The field is expanding rapidly, with 8 articles in 2021 and 106 in 2025 (90.79% annual growth rate). Science mapping (keyword co-occurrence, bibliographic coupling, and co-citation) indicates that the field is grounded in technology-acceptance and self-efficacy theory, and features convergence around generative AI, AI literacy, and ethical application. A thematic map places educational technology and ChatGPT among motor themes and generative AI among emerging themes. A complementary content analysis of all 208 articles yields six thematic categories, the largest being students’ AI literacy, perceptions, and learning with AI tools (25.0%), AI in health-professions and other disciplinary education (21.2%), and human–AI learning frameworks (19.2%). Ethics and responsible-AI terms appear in about 40% of the articles, adoption terms in 30%, and educator-preparedness terms in 36%. The results indicate that responsible AI use is a prominent strand of the field, and give guidelines for policy, institutional leadership, faculty development and future research. The findings inform AI-in-education scholarship and educational policy aligned with United Nations Sustainable Development Goal 4 (SDG 4).