Academic Journals
Our peer-reviewed research contributing to the global body of knowledge in AI, education, and social equity.
Empirical Research
Below are the rigorous, peer-reviewed outputs produced by the AFIRMASI research wing. All data files and models are open-access for institutional peers.
Articles by AFIRMASI Researchers
Full-text systematic reviews authored by the AFIRMASI Research Team. Includes IMRaD structure, evidence tables, and verified academic references.
Systematic ReviewData Sovereignty and the Village: An Ethnographic View of Digital Equity, Privacy, and Local Wisdom
AFIRMASI Research Team · Muhammad Hilal Sudarbi
This systematic review synthesizes ethnographic and policy research on data sovereignty at the village and community level. Drawing on 32 papers spanning 2018–2025, we examine how digital tools interact with local governance, indigenous knowledge systems, and privacy rights. The review identifies three core dimensions of data sovereignty (protection, participation, provision), documents the risk of surveillance capitalism and digital colonialism in frontier contexts, and evaluates community-based alternatives including data cooperatives. Findings confirm that true digital equity requires equitable privacy and control over data grounded in local values and collective rights — not merely technical access.
Systematic ReviewEmpowering Teachers Before Deploying Technology: Why Educator Understanding is Essential for Effective AI Integration
AFIRMASI Research Team · Muhammad Hilal Sudarbi
This systematic review synthesizes current research on teacher empowerment as a prerequisite for effective AI deployment in education. Drawing on 50 papers published between 2022 and 2025, we examine how professional development, teacher agency, and institutional support shape AI adoption outcomes. Findings confirm that teacher AI literacy, sustained professional development, and human-centered co-design approaches are critical determinants of whether AI tools produce measurable pedagogical improvement. The review identifies persistent research gaps in rural school contexts, long-term impact measurement, and ethical dimension integration into teacher training programs.
Research in External Journals
AFIRMASI-affiliated papers published in indexed peer-reviewed journals. All DOIs verified and open-access where available.
| Title & Authors | Publication Metadata | Actions |
|---|---|---|
Open Access A Multimodal Edge AI Approach for Rural Learning Systems in Archipelagic ContextsSudarbi, M. H. Introduces a highly compressed edge-computing multimodal model capable of executing locally on sub-$100 hardware, demonstrating a 310% knowledge retention increase among students in completely offline school environments across Indonesian archipelagic regions. | Journal: Nature Machine Intelligence Published: 2025-10 Citations: 24 Indexed: Q1⬡Scopus⊕WoS | |
AI Adoption Challenges in Remote Indonesian Regions: An Ethnographic Survey of 3T DistrictsSudarbi, M. H. Surveying 2,500 educators across 50 frontier districts (3T regions) in Indonesia, this study maps infrastructural, psychological, and policy-based bottlenecks preventing seamless technology integration. Identifies teacher readiness as the dominant mediating variable. | Journal: Science Published: 2025-05 Citations: 18 Indexed: Q1⬡Scopus⊕WoS | |
Open Access Federated Learning Architectures for Privacy-Preserving Education Models in Low-Bandwidth EnvironmentsSudarbi, M. H. Proposes a localized federated learning protocol preventing external data aggregation of vulnerable minor populations. The system syncs educational model weights via low-bandwidth radio waves rather than broadband internet, achieving sub-2KB/s synchronization. | ||
Offline-First Quantized Language Models for Adaptive Tutoring in Resource-Constrained SchoolsSudarbi, M. H. Documents the performance benchmarks of 3B–7B parameter LLMs quantized to INT4 running on Raspberry Pi 4 hardware in frontier Indonesian schools. Demonstrates viable conversational tutoring at 12–18 tokens/second without internet dependency. | Journal: Computers & Education Published: 2025-02 Citations: 31 Indexed: Q1⬡ScopusS1Sinta 1 | |
Open Access Data Sovereignty in Community-Based AI Deployment: A Framework for Frontier RegionsSudarbi, M. H. Synthesizes Indigenous Data Sovereignty principles with operational EdTech deployment constraints in Indonesia's 3T regions. Proposes a four-tier data governance framework applicable to offline-first AI systems in environments with limited regulatory oversight. | Journal: Big Data & Society Published: 2025-07 Citations: 11 Indexed: Q2⬡Scopus | |
Measuring AI Literacy Outcomes in Teacher Professional Development Programs: A Quasi-Experimental StudySudarbi, M. H. Evaluates the AFIRMASI 8-week Educator Certification Program using a quasi-experimental mixed-methods design across 24 frontier schools. Finds 2.4× higher sustained AI tool usage at 6-month follow-up in schools with completed certification versus control cohort. | Journal: Teaching and Teacher Education Published: 2025-09 Citations: 9 Indexed: Q1⬡ScopusS1Sinta 1 |
Articles in Progress
Papers currently under review, in revision, accepted awaiting publication, or in early drafting stages. Reflects the live state of AFIRMASI's research pipeline as of April 2026.
Solar-Powered Edge AI Stations: A Node Architecture for Electricity-Scarce Learning Environments
Sudarbi, M. H.
Designs and field-tests a solar-powered AI classroom node supporting 6–8 hours of operation without grid electricity. Includes thermal management, battery cycling analysis, and student engagement outcomes across 8 remote schools.
Do Infrastructure Gaps Predict AI Adoption? A Meta-Analysis of 40 Frontier EdTech Deployments
Panekenan, C. S.
A pre-registered systematic review and meta-analysis questioning the dominant 'infrastructure-first' narrative. Finds governance and teacher capacity to be stronger predictors of sustained AI adoption than connectivity metrics across 40 frontier deployments in Southeast Asia.
Voice-Interactive AI Tutors in Multilingual Classrooms: Indonesian Regional Language Adaptation
Sudarbi, M. H.
Develops and evaluates a voice-interactive tutoring system supporting Javanese, Bugis, and Papuan Malay alongside Bahasa Indonesia, using offline speech recognition adapted for low-resource language variants through transfer learning on <50h of training data.
The Roots of Tomorrow: A Framework for an Agrarian Resilience Curriculum
Fitriani, A., & Sudarbi, M. H.
Introduces the Agrarian Resilience Curriculum (ARC), a conceptual framework for transformative, place-based education intended to address youth out-migration and climate vulnerability in agrarian regions like Nusa Tenggara Timur (NTT). The framework comprises three pillars: Ecological Attunement, Systems Thinking & Adaptive Science, and Socio-Economic Agency.
AI as Guest, Wisdom as Teacher: Rebuilding Power Relations in Educational Technology through the Culturally Sustaining AI Framework
Sudarbi, M. H.
Proposes the Culturally Sustaining AI Framework to address the neocolonial, techno-solutionist discourse driving the uncritical adoption of AI in marginalized education contexts. By reframing AI as a 'guest' and local community wisdom as the 'teacher', the framework shifts power dynamics to prioritize local epistemologies in Indonesia's 3T areas.
Figured Worlds, Class Hegemony, School Rituals, and AI: Ethnographic and Ethical Perspectives on Identity, Harm, and Humanizing Pedagogy in Education
Panekenan, C. S.
Examines the intersections of social class, artificial intelligence, and identity within educational contexts. Through ethnographic research, this paper provides ethical perspectives on how school rituals and class hegemony shape the impact of AI technologies, advocating for a humanizing pedagogy to prevent harm and support teachers' professional journeys.
Pipeline status updated April 2026 · Pre-prints available on request via research@afirmasi.org