Robotics and Artificial Intelligence in Education: Transformations, Challenges, and Future Directions

Synthesis: White & Wu (2026) critically examine the integration of AI and robotics into education, arguing that while transformative potential exists at all levels, effective and equitable integration requires sustained structural investment, coherent policy, rigorous teacher preparation, and ethical practice. The review — grounded in peer-reviewed literature and structured around research trends, AI literacy, teaching roles, learning outcomes, STEM education, and ethics — finds that the field has moved faster than its evidence base. Research remains geographically concentrated, methodologically short-term, and insufficiently attentive to equity, human-centered design, and the broader social implications of automation in education.

Key Findings

1. Research remains geographically concentrated and short-term. Most empirical studies involve learners under 13 across periods of four weeks or fewer, concentrated in a small number of early-investing countries, which limits the generalizability of conclusions to diverse contexts.
2. AI and robotics can enhance learning, but gains are unevenly distributed. Evidence links them to improved personalized learning, Student Engagement, and STEM performance, yet infrastructure deficits, financial constraints, and faculty unpreparedness block equitable access — particularly in the Global South.
3. The dominant stance is augmentation, not replacement. The literature resists the idea that AI and robots can substitute for the relational, emotional, and cognitive dimensions of human teaching; the role of teachers is reframed and supported rather than removed.
4. AI literacy is an urgent but underdeveloped priority. Robotics-based tools can support AI literacy across educational levels, yet systematic, scalable, and equitable approaches remain rare outside well-resourced institutions.
5. Ethics and policy lag the pace of deployment. National AI policy strategies largely overlook the educational, ethical, and equity implications of AI, and structured frameworks for data privacy, algorithmic bias, and AI Governance remain nascent.
6. Perceptions of robots are positive but shaped by experience and misconception. Prior exposure and disciplinary context drive attitudes, and structured teacher education can correct Misconceptions about AI among future teachers.

Research Landscape and Trends

The review synthesizes a rapidly expanding but unevenly developed body of work on AI and robotics in education. Chu et al. (2022) analyzed SSCI journal articles from the Web of Science and found that Canada, Chile, and South Korea were early investors, with most studies targeting learners under 13, running four weeks or fewer, in physical school environments. Language and science dominated the application domains, with AI robots most often assuming the roles of tutor or tutee and learning performance serving as the most widely studied outcome. Later syntheses — Fombona et al. (2025, 132 articles) and Chen et al. (2023, structural topic modeling) — extend this picture, documenting the spread of AI robots and chatbots across early education, STEM, medical education, and language learning. Chen et al. (2023) nonetheless note the limited development of genuinely human-centered AI in educational robotics and call for greater learner involvement in robot design and testing. A bibliometric analysis by Lampropoulos (2025) of 361 documents characterizes social robots by their physical presence, anthropomorphic qualities, and affective and cognitive capabilities.

AI Literacy and Educational Robotics

A recurring theme is the need to build AI literacy in both students and teachers. The Robobo Project (Bellas et al., 2024), validated over six years, uses intelligent robotics to support formal AI literacy training across levels, teaching AI fundamentals such as perception, representation, reasoning, and learning. Eguchi (2021a, 2021b) argues that AI-powered educational robotics can advance computational thinking and AI literacy in K–12 settings aligned with the AI4K12 Five Big Ideas, and introduces CogBots — an Open Source, affordable AI-robotics tool developed with Google, CogLabs, and UNESCO for underserved communities. Kandlhofer et al. (2019) developed the European Driving License for Robots and Intelligent Systems, a competency-based, blended certification system for AI and robotics at the K–12 level. Experimentally, Hijón-Neira et al. (2024) found that preservice teachers using AI-generated assignments showed higher engagement, improved Problem Solving and algorithmic thinking, and greater confidence in teaching with educational robots.

Teaching Roles and Pedagogical Models

Transformation of the teacher's role is a central concern. Selwyn (2019) frames AI integration as a matter of values, judgments, and political choice rather than an inevitable process, and questions whether AI can replicate the social, emotional, and cognitive qualities of human teachers. Hrastinski et al. (2019) found practitioners saw greater potential for AI than for robots in supporting individualization, while also raising concerns about ethics and economic interests. Reviews by Alam (2021) and Mobo et al. (2025) trace how AI platforms and robotics enable instructors to improve the quality and efficiency of their work, customize curricula, and disrupt conventional pedagogical models, pointing toward AI-driven analytics, collaborative robots, virtual and augmented reality, and lifelong learning support. Across the literature the dominant position is complementarity: AI and robotics augment rather than supplant human teaching, freeing educators from routine tasks — but this promise only holds if teachers are adequately prepared, a dimension consistently identified as critical and underserved.

Learning Outcomes, Engagement, and STEM

Empirical evidence on learning outcomes and engagement appears across multiple studies. Anowar and Khatun (2024) report that AI-driven adaptive systems improve personalized learning and that robotic integration has increased student engagement and STEM performance, citing India's Atal Tinkering Labs as having reached over five million students. Abdujalilov (2024) found students taught with an AI- and robotics-focused methodology outperformed a control group. Vasou et al. (2024) surveyed 248 respondents in Cyprus and Greece, finding a strong correlation between educational attainment and appreciation of AI's role, alongside infrastructure and financial barriers. Salas-Pilco (2020) used a design-based research methodology to show how an integrated analytical framework yields more inclusive assessment of physical, social-emotional, and intellectual learning outcomes. In STEM, Malec (2001) notes robotics has been used since the late 1980s, yet benefits have remained difficult to isolate — a methodological challenge that persists. Loreggia (2024) finds technology-driven approaches create opportunities for engagement in STEAM subjects, while Ghofur (2025) documents improved engagement in Indonesian higher education tempered by infrastructure limits, financial constraints, and faculty training gaps.

Ethics, Equity, and Policy

The ethical and equity dimensions are the least adequately addressed. Schiff (2022) conducted a thematic analysis of 24 national AI policy strategies and found that the use of AI in education was largely absent from mainstream policy conversation, with priority given to education's instrumental role in producing an AI-ready workforce and scant attention to ethics. Adel (2024) maps challenges including data privacy, the digital divide, teacher and student readiness, and potential biases in AI-driven systems in the context of smart education. Virvou and Tsihrintzis (2024) propose the FEPER framework for the pedagogically effective and ethical use of AI tools, addressing interaction quality, data security, transparency, and trustworthiness. Akpomi et al. (2022) examine AI, robotics, and ICT in Nigerian education management, where ICT investment has not yet yielded commensurate outcomes. The authors stress that well-designed, resourced programs can produce gains, but that enthusiasm must be tempered by the risk that these technologies deepen existing divides between well- and under-resourced institutions, communities, and nations.

What this means for practice

• Instructors. Treat AI and robotics as augmentation, and budget teacher preparation alongside the hardware: the review finds the complementarity promise holds only where teachers are adequately prepared, a dimension it repeatedly identifies as critical and underserved.
• Instructors. Survey prior exposure and disciplinary context before rollout — both drove attitudes toward robots in the studies reviewed — and address misconceptions directly, since structured teacher education corrected them among future teachers.
• Instructors. Use robotics-based tools as a route into AI literacy, teaching perception, representation, reasoning, and learning as the Robobo Project does, rather than treating them as an add-on for STEM performance alone.
• Instructors. Evaluate beyond learning performance: the review notes that performance metrics dominate the literature and urges attention to social-emotional, physical, and civic outcomes, with learners involved in robot design and testing rather than only assessed by it.
• Instructors. Audit the equity ledger before scaling: the review warns that these technologies can deepen divides between well- and under-resourced institutions, communities, and nations, so programs should be checked for who actually gains access.

Limitations

• The paper is a discursive review and critical opinion piece over a predefined corpus: all sources were supplied as a fixed reference set and no additional sources were introduced, so it cannot claim the coverage or reproducibility of a systematic search.
• It reports no primary data and no meta-analysis; field-level trend claims — for instance that most studies run four weeks or fewer with learners under 13 — rest on secondary reviews such as Chu et al. (2022) rather than on the authors' own coding of primary studies.
• Its conclusions inherit the limits the review itself documents for the field — geographically concentrated, short-term, performance-metric-heavy research — so the recommendations describe what the current evidence base can support, not measured effects for any specific implementation.

Citation

White, A. R., & Wu, Z. (2026). Robotics and Artificial Intelligence in Education: Transformations, Challenges, and Future Directions. EdArXiv. doi:10.35542/osf.io/ebyhj_v1.