Understanding Geography Teachers' Assessment Challenges and Needs for Artificial Intelligence Supported Assessment Systems: Insights from Indonesia
This mixed-methods study of Indonesian geography teachers identifies their specific assessment challenges and pedagogical needs to propose a teacher-centred conceptual framework that guides the design of effective, context-responsive AI-supported assessment systems.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In classrooms around the world, the way teachers check what students have learned is undergoing a quiet but profound shift. For decades, assessment was often a final event: a test at the end of a unit that measured what a student knew at that specific moment. Today, educators are increasingly trying to turn assessment into a continuous conversation, a tool that helps students learn as they go rather than just judging them after the fact. This approach, often called "assessment for learning," requires teachers to provide quick, personalized feedback and to track every student's progress in real time. However, doing this well is incredibly difficult. A single teacher might manage dozens of students, each with unique gaps in understanding, and the sheer volume of work needed to give meaningful, individualized guidance often outstrips the hours in a school day. As schools look to technology to solve this problem, a new question has emerged: can artificial intelligence, the computer systems capable of learning and making decisions, actually help teachers manage this workload without losing the human touch that makes education work?
A team of researchers in Indonesia set out to answer this question by listening directly to the people who would be using the technology: geography teachers. They focused their study on Surakarta, a city known for its digital innovation, where teachers are already navigating a new national curriculum that emphasizes deep, critical thinking over simple memorization. The researchers wanted to understand not just whether these teachers were ready to use artificial intelligence, but specifically what problems they hoped it would solve. They gathered a group of forty geography teachers from high schools in the city and asked them to fill out a detailed survey about their daily challenges, their feelings toward technology, and what features they would want in a smart assessment system. To get a deeper understanding, the researchers then brought these same teachers together for group discussions, allowing them to explain the stories behind their survey answers.
The results revealed a clear and surprising picture. The teachers were not skeptical of artificial intelligence; in fact, they were eager to adopt it. However, their enthusiasm was not driven by a fascination with high-tech gadgets or a desire to automate everything. Instead, their interest was rooted in very practical, everyday struggles. The teachers reported that while they were good at giving tests, they often lacked the time to provide the kind of detailed, immediate feedback that helps a student improve. They found it difficult to monitor the progress of every single student in a large class, and they struggled to help students learn how to evaluate their own work. The study found that the teachers' desire for artificial intelligence was directly linked to these specific pain points. The more a teacher struggled with providing timely feedback or tracking individual progress, the stronger their need for a system that could help.
When the researchers asked the teachers what they wanted from an artificial intelligence system, the answers were specific and grounded in the reality of the classroom. The teachers did not want a machine that would simply grade papers and hand back a score. They wanted a partner that could generate a variety of questions tailored to the curriculum, analyze a student's answers to spot exactly where their thinking went wrong, and then immediately suggest what material the student should review next. They envisioned a system that could visualize a student's learning journey over time, showing a teacher at a glance which students were falling behind and which were ready for a challenge. Crucially, the teachers emphasized that the system needed to be easy to use and must not add to their already heavy administrative burden. They saw artificial intelligence not as a replacement for their professional judgment, but as a tool that could handle the repetitive tasks of data collection and initial analysis, freeing them up to focus on the human side of teaching: interpreting the data, guiding the students, and making the final decisions about how to help each learner.
The study concluded that for artificial intelligence to succeed in schools, especially in developing countries where resources can be limited, it must be designed around the needs of the teacher rather than just the capabilities of the technology. The researchers proposed a new framework where the starting point for building these systems is the specific challenge a teacher faces in the classroom. If a teacher is overwhelmed by the time it takes to give feedback, the technology should be built to solve that specific problem. The findings suggest that the future of educational technology lies in this kind of teacher-centered design. By focusing on the real, daily hurdles teachers face, developers can create systems that do more than just process data; they can become genuine partners in the learning process, helping to turn the complex task of assessment into a powerful engine for student growth.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.