CHATGPT EN LA ENSEÑANZA INICIAL DE PROGRAMACIÓN: DINÁMICAS EMERGENTES EN UN ECOSISTEMA DE APRENDIZAJE COMPLEJO

Abstract

Learning to program for the first time often sparks curiosity and, at the same time, a certain amount of tension. For many university students, encountering code means entering uncharted territory where logic operates according to different rules and mistakes are a constant companion. Some approach it with enthusiasm; others, with caution. What almost all share is the feeling of undergoing a learning process unlike anything they are used to. In the midst of this process, generative artificial intelligence tools, such as ChatGPT, have begun to be incorporated. Their presence changes the dynamic: they allow students to ask questions without feeling judged, review an idea before implementing it, or delve into a concept that wasn't clear in class. For some students, it was a kind of silent guide. For others, a resource they only turned to when they were clearly stuck. This study explores that experience in students who began programming with Python. The analysis is grounded in complexity theory, which understands learning as a living process, with advances, setbacks, discoveries, and moments of doubt. There wasn't just one way to use the tool. There were different learning paths: some marked by dependence, others by increasing autonomy. And, in several cases, something changed in the relationship with error: it ceased to be felt as failure and began to be seen as part of the learning process. The final reflections address the pedagogical and ethical implications of integrating generative systems into introductory programming courses, recognizing that the use of these tools not only modifies the practice but also the way students relate to learning itself.