Ethical issues at the intersection of artificial intelligence and sustainability: A systematic review and its implications for science, mathematics, and technology education

Abstract

Artificial intelligence (AI) is increasingly presented as a tool for sustainable development, yet the ethical costs of this role remain scattered across separate bodies of literature. This systematic review aims to identify the ethical issues arising at the intersection of AI, ethics, and sustainability. Following the preferred reporting items for systematic reviews and meta-analyses 2020 statement, Scopus and Web of Science were searched in February 2026, yielding 5,794 records. Screening, eligibility assessment and quality appraisal were carried out by two researchers, and 74 studies were included. Verbatim passages addressing the intersection were extracted from each study and examined using a hybrid thematic approach, resulting in 921 coded passages. Five themes were identified: algorithmic bias and unequal AI outcomes; privacy, surveillance, and human autonomy; transparency, accountability, and epistemic integrity; AI governance and corporate responsibility; and the ecological ethics of AI systems. Concerns regarding transparency and accountability accounted for the largest share of passages (30.9%), while ecological concerns were voiced across all five themes rather than in a separate discussion. The themes were mapped onto the environmental, social, and governance dimensions of sustainability and onto the United Nations sustainable development goals. Transparency and accountability thus emerge as a structural prerequisite rather than a procedural add-on for sustainable AI. Educational research is markedly underrepresented at this intersection (5 of 74 studies); the implications of the five themes for science, mathematics, and technology education are therefore developed in detail, from carbon- and water-footprint modelling tasks to data-justice and AI-literacy work in the classroom.