Criterios de corrección de exámenes tradicionales de Física y Química: docentes en formación frente a la inteligencia artificial

Abstract

The evaluation of Physics and Chemistry in secondary school classrooms continues to rely predominantly on traditional exam-based assessments. Despite advances in research on assessment methods and the benefits of formative evaluation, resistance to changing the types of assessment instruments remains strong. This study aimed to compare the correction and grading criteria used by preservice Physics and Chemistry teachers with those integrated into Artificial Intelligence (AI) tools, specifically ChatGPT and Gemini. A total of 105 secondary education preservice teachers participated in the study. The results reveal a lack of reliability due to inconsistencies in the application of correction criteria, with no significant differences compared to those employed by AI tools. However, the AI demonstrated greater rigor in applying these criteria. The findings encourage diversifying the types of instruments used to assess the teaching and learning process in secondary Physics and Chemistry and suggest leveraging AI for grading when traditional exams are maintained.