AI evidence extraction

AI-enhanced STEM education: A bibliometric study of research trends toward achieving sustainable development goals

European Journal of STEM Education · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

6/10

Relevance

0/4

Quality (LMQS)

Evidence

Citations

6.32

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.20897/ejsteme/18190

Methodology & findings

Study design

Bibliometric analysis using PRISMA method.

Main result

The findings reveal a significant increase in studies on the use of AI in STEM education, with "the dominant themes being personalized learning, machine learning, and, more recently, generative AI, such as ChatGPT." The study also found that "integrating AI into STEM education can significantly contribute to achieving SDG 4 (quality education)" while supporting workforce development initiatives aligned with SDG 8.

Research paradigm

positivist

Author conclusions

The authors conclude that "This research informs the strategic planning of AI-based applications in STEM education worldwide and provides insights into policymakers, scientists, and educators, while remaining aligned with the Sustainable Development Goals." They emphasize that the research addresses both educational quality and workforce development needs through AI integration.

Risk of bias

Geographic bias in research representation (limited to USA, China, and Singapore); Language bias (Scopus database search may exclude non-English publications); Publication bias (only published documents included); Unbalanced collaboration networks among countries; Publication bias (limited to Scopus database); Geographic bias (dominated by USA, China, and Singapore research); Language bias (likely English-language documents only); Selection bias (PRISMA method applied but database limited to Scopus); Selection bias: Document selection limited to Scopus Database only, potentially excluding relevant literature from other databases; Geographic bias: Research landscape dominated by wealthy nations (USA, China, Singapore) with unbalanced collaboration networks; Publication bias: May favor published, peer-reviewed literature over grey literature; Language bias: Scopus indexing typically favors English-language publications

Limitations

The authors note that "SDG 10 (reduced inequalities) is a point of concern raised by the limited extent of research on a few nations
This is due to the possibility that economically and educationally diverse international settings might not be helped by AI-driven solutions and research agendas from rich contexts, potentially triggering increasing economic and educational inequalities." Additionally, they identify that "Different countries, such as the USA, China, and Singapore, lead the research landscape but have unbalanced collaboration networks."

Open questions raised

The study identifies that research on AI in STEM education is concentrated in economically developed nations (USA, China, Singapore), with limited research in educationally and economically diverse international settings. There is concern that AI-driven solutions from wealthy contexts may not adequately address needs in developing regions, potentially exacerbating educational and economic inequalities.
The study identifies unbalanced collaboration networks among leading countries and highlights the need for more research in economically and educationally diverse international settings to prevent widening inequalities through AI-driven education solutions.
The authors identify that limited research exists on AI applications in STEM education in economically and educationally diverse international settings, and that unbalanced collaboration networks among leading countries (USA, China, Singapore) indicate a research gap regarding global equitable access to AI-driven educational solutions and SDG 10 (reduced inequalities).

Data: not_statedCode: not_statedExtracted from: pdfAgreement 73%

Explore related topics

Scholarly Infrastructure Literature Review