Revisiting a Teaching Sequence on the Topic of Electrolysis: A Comparative Study with the Use of Artificial Intelligence

Abstract

The literature has reported the growing use of Artificial Intelligence (AI) tools for educational purposes. This article aims to discuss the potentialities and limitations of AI in lesson planning and teacher training in chemistry and science subjects. The study compares three high school Teaching Sequences (TSs) on electrolysis, developed with and without AI’s assistance. The first, H-TS, was crafted within a Brazilian university’s Chemistry Teaching program, without AI. The other two, GPT-3.5-TS and GPT-4-TS, were independently created using GPT-3.5 and GPT-4, maintaining consistent lesson numbers and durations, and adhering to the same theoretical and methodological framework, including the use of a simulator. Findings indicate a marked improvement in TSs developed with GPT-4 over GPT-3.5. Nonetheless, both AIs exhibited limitations such as shallow reasoning and repetitive use of prompt terms, lacking in-depth didactic-pedagogical elaboration. They produced highly prescriptive sequences, potentially reinforcing misconceptions, and displayed a relative disregard for the educational context─aspects present to a lesser extent in H-TS. On a positive note, AIs demonstrated strong organizational skills, relatively adequate framework mapping, and diverse instructional and assessment strategies. The study suggests a moderate and formative use of AIs in educational contexts. It underscores the necessity for teachers to critically engage with AI-generated content and iteratively use AI to refine sequences, ensuring these do not become final products. Future research should explore AI’s practical application in TS development and its impact on preservice teacher training, aiming to critically assess and address the complexities and implications of AI-generated TSs.