Implementation of AI tools (NotebookLM) for technical training on regulations related to hazardous waste

Abstract

Sustainable management of edaphic resources and proper disposal of organic waste (biosolids) constitute critical challenges for modern agriculture and environmental engineering. The technical complexity of regulations such as NOM-004-SEMARNAT-2002 and NOM-138-SEMARNAT/SSA1-2012 creates significant learning barriers for future professionals responsible for their application. This study investigates the efficacy of a pedagogical strategy mediated by Generative Artificial Intelligence (NotebookLM) to transform dense regulatory texts into audio formats (podcasts), facilitating the assimilation of remediation and waste management protocols. Under a quasi-experimental design at the Universidad Juárez Autónoma de Tabasco, the performance of an experimental group ($n=22$) was compared against a controlled historical baseline. Rote memorization assessment was replaced by the resolution of practical soil contamination scenarios. Results indicate a statistically significant improvement ($\chi^2 = 9.84, P < 0.01$), reducing the failure rate from 60% to 18.2%. Furthermore, a notable increase in diagnostic and remediation proposal competencies was evidenced. It is concluded that AI audio generation technology reduces extrinsic cognitive load, allowing for deeper appropriation of agricultural and environmental technical regulations.