Human Experts' Evaluation of Generative AI for Contextualizing STEAM Education in the Global South

📝 Abstract

STEAM education in many parts of the Global South remains abstract and weakly connected to learners sociocultural realities. This study examines how human experts evaluate the capacity of Generative AI (GenAI) to contextualize STEAM instruction in these settings. Using a convergent mixed-methods design grounded in human-centered and culturally responsive pedagogy, four STEAM education experts reviewed standardized Ghana NaCCA lesson plans and GenAI-generated lessons created with a customized Culturally Responsive Lesson Planner (CRLP). Quantitative data were collected with a validated 25-item Culturally Responsive Pedagogy Rubric assessing bias awareness, cultural representation, contextual relevance, linguistic responsiveness, and teacher agency. Qualitative reflections provided additional insight into the pedagogical and cultural dynamics of each lesson. Findings show that GenAI, especially through the CRLP, improved connections between abstract standards and learners lived experiences. Teacher Agency was the strongest domain, while Cultural Representation was the weakest. CRLP-generated lessons were rated as more culturally grounded and pedagogically engaging. However, GenAI struggled to represent Ghana’s cultural diversity, often producing surface-level references, especially in Mathematics and Computing. Experts stressed the need for teacher mediation, community input, and culturally informed refinement of AI outputs. Future work should involve classroom trials, broader expert participation, and fine-tuning with Indigenous corpora.

💡 Analysis

STEAM education in many parts of the Global South remains abstract and weakly connected to learners sociocultural realities. This study examines how human experts evaluate the capacity of Generative AI (GenAI) to contextualize STEAM instruction in these settings. Using a convergent mixed-methods design grounded in human-centered and culturally responsive pedagogy, four STEAM education experts reviewed standardized Ghana NaCCA lesson plans and GenAI-generated lessons created with a customized Culturally Responsive Lesson Planner (CRLP). Quantitative data were collected with a validated 25-item Culturally Responsive Pedagogy Rubric assessing bias awareness, cultural representation, contextual relevance, linguistic responsiveness, and teacher agency. Qualitative reflections provided additional insight into the pedagogical and cultural dynamics of each lesson. Findings show that GenAI, especially through the CRLP, improved connections between abstract standards and learners lived experiences. Teacher Agency was the strongest domain, while Cultural Representation was the weakest. CRLP-generated lessons were rated as more culturally grounded and pedagogically engaging. However, GenAI struggled to represent Ghana’s cultural diversity, often producing surface-level references, especially in Mathematics and Computing. Experts stressed the need for teacher mediation, community input, and culturally informed refinement of AI outputs. Future work should involve classroom trials, broader expert participation, and fine-tuning with Indigenous corpora.

📄 Content

Contextualization in STEAM (Science, Technology, Engineering, Arts, and Mathematics) education has become an urgent pedagogical priority, particularly within the Global South, where educational content often fails to reflect learners’ sociocultural realities (Ngman-Wara (Ngman-Wara, 2015;Nsengimana et al., 2020). In Ghana, for instance, Ngman-Wara (2015) revealed that junior high school science teachers had limited knowledge of contextualized science instruction, despite its alignment with national educational reforms. The study highlighted a significant disconnect between curriculum standards and actual classroom practice, where science was often taught devoid of local examples, analogies, or linguistic relevance (Ngman-Wara, 2015). This finding underscores a broader challenge across the Global South: while policy documents advocate for culturally and locally responsive pedagogy, teachers frequently lack the training, resources, and tools to implement it effectively, especially within STEAM domains where globalized content dominates (Nsengimana et al., 2020;Thibeault-Orsi, 2022).

The intensifying workload of teachers compounds this challenge (Do Minh et al., 2021). Lesson planning, a critical yet time-intensive task, often competes with grading, administration, and extracurricular duties (Sakamoto et al., 2024;Şimşek, 2025;Zheng & Stewart, 2024). Consequently, educators are increasingly turning to Generative AI (GenAI) tools to streamline lesson development. Studies such as van den Berg and du Plessis (2023) and Kerr and Kim (2025) show that GenAI can support lesson structure, resource generation, and language editing. However, these same studies caution that AI-generated lessons often lack context sensitivity, frequently omitting cultural and linguistic features essential to equitable STEAM education. As Hamouda et al. (2025) observed in their cross-African computing education review, efforts to localize content remain scattered and under-theorized, revealing the systemic difficulty of contextualizing STEAM teaching across diverse African classrooms.

One recent intervention addressing this gap is the Culturally Responsive Lesson Planner (CRLP), developed by Nyaaba and Zhai (2025). Grounded in culturally responsive pedagogy (CRP), their semi-interactive GenAI tool prompts teachers to embed local language, practices, and examples directly into AI-generated lesson plans (Nyaaba & Zhai, 2025;Nyaaba et al., 2024). Their findings showed that CRLP outputs outperformed generic GPT outputs in curriculum alignment, cultural relevance, and language use. However, their study stops short of comparing these CRLPs directly with human-created or standards-based lesson plans, a gap this study seeks to address. In this study, we seek to advance this line of inquiry by conducting a comparative analysis between GenAI-generated and curriculum-based (human-developed) lesson plans for STEAM education in Ghana. We employed human experts to generate STEAM lesson plans using Nyaaba and Zhai (2025) CRLP, evaluated them, and reflected on both the CRLP-generated lesson plans and the National Council for Curriculum and Assessment (NaCCA) lesson plans to determine their relative effectiveness in promoting culturally grounded and pedagogically sound instruction in STEAM education. Through this process, the study aims to uncover how GenAIgenerated lessons can complement or enhance existing curricular materials and contribute to the advancement of equitable and contextually relevant STEAM education in the Global South. We emphasized human pedagogical judgment to ask the following questions.

1. How do experts evaluate the cultural responsiveness and pedagogical quality of GenAIgenerated lesson plans? 2. In what ways do experts compare GenAI-generated lesson plans to standards-based (NaCCA) lesson plans?

Theoretical Framework

This study is grounded in two complementary theoretical perspectives: Culturally Responsive Pedagogy (CRP) and Human-Centered Artificial Intelligence (HCAI), situated within the broader concept of Human-AI Collaboration (Gay, 2015;Gay & Howard, 2000;Güvel et al., 2025). These frameworks provide the foundation for understanding how GenAI-generated lesson plans can align with local cultural and pedagogical contexts and how expert reflection can guide the responsible use of AI in education. This study adapts the AI and Culturally Responsive Assessment Framework developed by Nyaaba et al. (2024) as its theoretical foundation.

The framework integrates the principles of Culturally Responsive Pedagogy (CRP) with the evolving capabilities of GenAI to explore how GenAI can meaningfully support culturally grounded assessment. Their Nyaaba et al. (2024) framework embeds the classical works of Ladson-Billings (1995) and Gay (2018) to situate culture as a dynamic force that shapes both teaching and learning while matching the potentials of GenAI. The framework rests on five cultural tenets: Indigenous language, Indigenous kno

This content is AI-processed based on ArXiv data.