This crucial finding was revealed in a new 2026 study conducted by Agustinus Sembiring and Theresia Herlina, two researchers from Pradita University, Tangerang. Released in the May 2026 edition of the Indonesian Journal of Advanced Research (IJAR), the study provides fresh empirical evidence that utilizing smart technology can bridge the gap in educational quality within rural or semi-rural areas, particularly in digital-based subjects like coding.
This milestone offers a breath of fresh air for the primary education sector, which has long been dominated by conventional teaching methods due to a lack of devices and network connectivity. The experimental results demonstrated that students who learned with the assistance of AI experienced an academic score leap that was twice as high as those who studied using standard methods without smart technological support.
The Technological Gap in Primary Education
Over the past few years, the rapid advancement of machine learning and generative AI has dominated various sectors, including the education sector. However, the implementation of this technology at the elementary school level remains highly unequal and unevenly distributed. Most previous academic studies focused heavily on urban environments or universities equipped with high-speed internet and modern gadgets.
A very different condition is observed in rural elementary schools, such as those in the Karo Regency of North Sumatra. In these areas, the majority of learning activities still rely on traditional, teacher-centered methods. While some educators have independently started using AI to draft lesson plans or manage classroom administration, integrating AI directly into live classroom learning interactions remains exceptionally rare.
Recognizing this research gap, Agustinus Sembiring and Theresia Herlina conducted field experiments to see if AI could remain highly effective even when deployed in an ecosystem with strict technological constraints. They focused on fifth-grade students—a demographic deemed to possess the basic digital adaptability required to engage with interactive learning platforms.
Simple Yet Rigorous Methodology
To gather valid and objective data, the research team from Pradita University implemented a quasi-experimental design. The study involved a purposive sample of 130 fifth-grade students from a private elementary school in Karo Regency.
The students were divided fairly into two major groups:
- Group A (AI-Based): The experimental group that received learning materials and evaluations through an interactive questionnaire powered by an adaptive AI system.
- Group B (Non-AI): The control group that followed conventional learning sessions using textbooks and standard whiteboards without any smart applications.
Before the experiment began, both groups took a pretest to measure their baseline capabilities. After the learning period concluded, both groups were re-evaluated using a posttest alongside the Evaluation of Learning Quality and Output (EKOP) questionnaire, which measures student perceptions of teaching performance, motivation, and classroom environment.
A Massive Leap in Grades: The AI Group Excels Unconditionally
The statistical data analysis revealed a striking contrast between the two student groups. The students who learned using AI technology achieved extraordinary grade improvements.
The average test score for the AI group surged from 56.65 on the pretest to 78.85 on the posttest. This represents an average score growth (gain mean) of 22.20 points.
Conversely, the conventional group without AI only experienced a minor increase, moving from an initial average of 57.15 to 67.60 on the final exam, yielding a gain of just 10.45 points. Mathematically, the learning growth of students supported by AI was more than double that of their peers learning through traditional methods.
Through a comparative statistical test (independent sample t-test), researchers obtained a highly robust significance value (t = 15.15; p < 0.001). This metric confirms that the AI group's superiority was not a matter of chance, but a direct result of the positive impact of adaptive technology.
Beyond test scores, data from the EKOP questionnaire showed that the AI group scored the highest across all indicators of learning process quality and output. For instance, in teaching performance and interactivity, the AI group scored an average of 4.25 out of 5, far outperforming the non-AI group, which only scored 3.34. The AI technology was credited with successfully boosting active student engagement, sparking curiosity, and fostering critical thinking and problem-solving skills from an early age.
Broad Implications for Public Policy and Education
The success of the experiment in Karo Regency carries profound implications for the future of education. This study dismantles the assumption that advanced technologies like AI are only viable in elite urban schools. AI adoption has proven to be an instant solution for delivering a "personal tutor" experience to children in underfunded facilities, helping teachers map out student learning difficulties automatically and in real time.
For public policymakers and educational boards, the study provides a strong recommendation: school digitalization programs must expand beyond distributing laptops to include providing child-friendly, AI-driven adaptive software. Teachers in rural areas need comprehensive training to seamlessly integrate artificial intelligence, enabling them to personalize educational materials based on each child's learning pace.
Researcher Profiles
Agustinus Sembiring, S.Kom., M.T. is a lecturer and researcher at Pradita University, Tangerang. He specializes in Educational Technology, Artificial Intelligence, and the development of digital learning media. His research portfolio focuses on leveraging adaptive platforms to enhance literacy among children in tech-limited areas.
Theresia Herlina, S.Pd., M.Pd. is an academic at Pradita University specializing in Curriculum, Educational Evaluation, and Primary School Teaching Methods. She is actively researching the optimization of technology-based classroom management to support 21st-century skills development.
Official Research Source:
Journal Article Title: The Use of AI in Education to Improve Elementary School Students' Learning Outcomes
Authors: Agustinus Sembiring & Theresia Herlina (Pradita University)
Journal Name: Indonesian Journal of Advanced Research (IJAR), Volume 5, Number 5, Year 2026, Pages 733-746.
Digital Object Identifier (DOI): https://doi.org/10.55927/ijar.v5i5.16539
0 Komentar