Metrology and Artificial Intelligence in Analytical Chemistry Education without Laboratory Infrastructure

Abstract

This study investigated, within the context of Analytical Chemistry education, the potential of a digital experimental activity grounded in metrological principles and mediated by artificial intelligence (AI) to promote meaningful learning. The research involved ten groups of undergraduate students enrolled in an Analytical Chemistry Instrumentation course at the Federal University of Piauí (UFPI), in a setting characterized by the absence of conventional laboratory infrastructure, which required the adoption of an accessible, data-centered experimental approach. Students performed indirect measurements of human growth using standardized digital images, applying internal standard calibration to convert pixel measurements into centimeters, organizing data into tables, constructing graphs, and developing linear regression models using the least squares method. Measurement quality was assessed through relative error percentage, with an a priori didactic acceptability criterion of ER ≤ 3%, explicitly addressing repeatability and metrological reliability. AI tools were integrated as supportive resources for verifying calculations, checking statistical consistency, generating graphical representations, and identifying inconsistencies, always under instructor mediation and without replacing students’ analytical reasoning. Quantitative results showed high consistency among groups that met the metrological quality criterion, with a mean estimated height of 142.9 cm, a standard deviation of 9.9 cm, and an average coefficient of determination of 0.988, indicating strong linear correlation. Qualitative findings demonstrated students’ conceptual understanding of calibration, measurement error, and reliability, as evidenced by written justifications, recognition of random and systematic errors, and interpretation of regression parameters. Comparative analysis of estimation approaches and contextualization of results as analytical references supported critical discussion of model assumptions, extrapolation limits, and the impact of measurement quality on result robustness. Overall, the findings indicate that integrating digital experimentation, metrology, and artificial intelligence constitutes a viable, accessible, and replicable strategy for teaching Analytical Chemistry in public universities, enhancing students’ data analysis skills, critical thinking, and evidence-based reasoning in resource-limited educational environments.