Abstract

Determining the level of cognitive demand in school tasks is a significant challenge, marked by the lack of consensus among evaluators when assigning cognitive complexity in different areas of knowledge. In view of this problem, we propose a methodology based on the Theory of Registers of Semiotic Representation to guide the assignment of cognitive demand in mathematical tasks, specifically focused on the quadratic function. We have used ChatGPT to implement this methodology, applying it to mathematical tasks extracted from school textbooks. The results show that the analysis methodology proposed for the classification of these tasks is simple to implement with ChatGPT and with the experts. Moreover, on the classification of the 3 tasks analyzed, the experts evaluated ChatGPT’s performance well with respect to 2 of them, and on a third one there were discrepancies.