Abstract
In many classrooms across the Caribbean, end-of-term assessments continue to be administered using paper-based examination scripts. Once the scripts are stored and the final scores are recorded, the hard copies are typically archived or discarded. This practice often results in a significant loss of valuable data that could otherwise inform formative feedback, diagnostic insight, and data-informed instructional planning. This paper introduces an AI-supported system, specifically built using a GPT-based large language model, optical character recognition (OCR), and spreadsheet automation techniques. The system transforms static handwritten exam scripts into dynamic, data-driven dashboards that provide teachers with actionable insight. The system performs three core functions: (1) it uses OCR to scan and extract students’ handwritten responses; (2) it processes these into structured, Excel-compatible datasets using a rule-based logic engine; and (3) it connects this data to a Google Sheets dashboard powered by automation scripts. The research illustrates the system’s functionality using anonymised data from a primary school classroom in Barbados. This study illustrates how the tool is applied in practice and how it can be deployed to enhance assessment practices in contexts with limited digital capacity. Beyond the technical demonstration, the paper explores ethical considerations associated with the use of AI in educational settings. Emphasis is placed on protecting student privacy through the use of anonymised codes and ensuring that teachers remain central to the interpretation of data. The potential risks related to bias, surveillance, and reliance on algorithms is also addressed.
Presenters
Grace Anne JackmanLecturer, School of Education, The University of the West Indies, Cave Hill Campus, Barbados
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
KEYWORDS
ASSESSMENT MONITORING, ARTIFICIAL INTELLIGENCE, GPT-BASED TOOLS, OCR, PRIMARY EDUCATION
