Artificial intelligence in K-12 education: An umbrella review

University college of education (with international co-author affiliation)

Illinois, USA

Author team umbrella synthesis and quality rubric application

• In-school (K–12)

Umbrella review of systematic reviews with framework building and quality assessment

Corpus of 102 systematic reviews (AIEd K-12 scope)

N/A — secondary synthesis of review literature

Spans all AI modalities represented in underlying reviews

• Dependent on quality and transparency of included reviews
• Fast-moving field may outpace review snapshots
• K-12 focus may omit adjacent informal learning reviews
• English-language and database biases possible

K-12 students and teachers as objects of reviewed AIEd research

Heterogeneous across reviewed contexts

Varies widely across reviews

• Map interconnected AIEd research topics for compulsory schooling
• Synthesize multi-domain innovations and challenges
• Flag underrepresentation of AI education/literacy and theoretical advancement

• Offer reusable AIEd Review Framework for future reviews
• Provide umbrella-review quality rubric for methodology improvement

• Not a primary-study meta-analysis of student learning effects
• Not a single-intervention evaluation
• Not exhaustive of grey literature unless captured by included reviews

Meta-level: aggregates tutors, recommender systems, surveillance, assistants, assessment AIs, etc.

• Tutor
• Automation tool
• Evaluator

Review-level English synthesis

• Characterizes how AI supports instruction, personalization, and operations at scale in reviewed work
• Highlights ethics and systemic barriers recurring across reviews

• Umbrella conclusions should not overstate evidence beyond constituent reviews
• Equity, surveillance, and data governance remain cross-cutting risks
• AI literacy for teachers/students needs dedicated review growth

• Collect systematic reviews on AIEd in K-12
• Chart characteristics and build framework of review foci
• Synthesize findings and challenges across bodies of work
• Assess review quality with rubric and identify methodological gaps

• Authors synthesize human-reported review claims; no automated reviewer in study

• Framework acts as navigational scaffold for future researchers

• AI applications in reviews outpace AI literacy and deep theory development reviews
• Many reviews apply frameworks descriptively without empirical extension
• Common weaknesses in data management, extraction transparency, and analysis rigor across reviews

• Introduces umbrella-specific quality rubric beyond PRISMA alone
• Explicitly quantifies review corpus (102) for K-12 AIEd

• Provides macro-map of AIEd innovation clusters in K-12

• Strengthens argument for more reviews on AI education/literacy and theory

Offers actionable navigation for researchers and signals where practitioner guidance remains thin at meta level.

• Synthesis must track surveillance and privacy-heavy AI uses highlighted across reviews
• Equity implications when AI supports sorting, prediction, or monitoring students
• Transparent reporting for future umbrella reviews using provided rubric
• Avoid policy prescriptions beyond strength of underlying reviews

• Activity documentation
• Practitioner observation

• Targeted systematic reviews on under-mapped AI literacy topics
• Primary studies testing theories referenced only descriptively in prior reviews

• AI in education meta-research
• K-12 policy
• Research methodology for reviews

• Completed

K-12

School systems (synthesized)

Multi-role AIEd

N/A (meta)

Varies

High in many underlying applications