ChatGPT-4 and DeepSeek-V3 in STEM Education: Evaluating AI-Generated Activities for an Industrial Engineering Course (#1191)

Authors

Villalobos Zamora, Miguel Angel

Abstract

The advancement of Artificial Intelligence (AI) has led to the development of large language models (LLMs), which are currently used to streamline processes in various fields. It is essential to understand their potential for implementation in education. This study evaluates the quality and integration of the STEM methodology in learning activities generated by ChatGPT-4o (OpenAI) and DeepSeek-V3 (DeepSeekAI) for the course C¸ omputer Tools II¨ın the Industrial Engineering program at Universidad Fidelitas, Costa Rica, in 2025. A quantitative research approach was used, employing ´ rubrics to assess the AI-generated responses. The results indicate that both LLMs successfully incorporate the STEM methodology, with ChatGPT-4o excelling in interdisciplinary integration and the promotion of critical thinking, while DeepSeek-V3 demonstrated strengths in source utilization and applicability. This research provides insights into the use of generative AI in education, highlighting its benefits and limitations in the design of educational activities.