Abstract:

This educational innovation examines Artificial Intelligence (AI) integration into competency development within a Food Engineering program. Focusing on the "Food Design" sub-competency, it evaluates the TechWizard tool's impact on product formulation under standardized technical specifications. The objective was to strengthen the teaching-learning process by utilizing predictive technology to streamline formulation and enhance design accuracy.

The intervention’s relevance lies in transitioning from empirical methods to predictive modeling. By employing AI-assisted reverse engineering, the tool processes nutritional data from commercial products in conjunction with official food composition databases to generate precise formulations and simulations prior to laboratory work. This approach integrates critical variables—international regulatory constraints, costs, and market demands—directly into the initial design phase. Consequently, AI optimizes workflow, reduces data errors, and minimizes experimental resource consumption, fostering autonomous and strategically grounded decision-making.

The research followed a quasi-experimental design using non-probabilistic convenience sampling with seventh-semester students (August-December 2025), compared against previous cohorts as a control group. Beyond technical performance, usability was measured using the User Experience Questionnaire Short (UEQ-S), following Sanchis Font et al. (2018) validation criteria for pragmatic and hedonic qualities. Student engagement was analyzed through the UWES-S scale (Schaufeli et al., 2002), assessing psychological commitment through its Vigor, Dedication, and Absorption dimensions.

Results demonstrated that AI-assisted reverse engineering facilitates rapid access to complex food composition data, significantly reducing laboratory residence time. By prioritizing high-probability simulations, students shifted focus from manual calculations to critical validations and technical-regulatory compliance. However, the UEQ-S identified a significant technical barrier; qualitative feedback described the interface as "unfriendly." Despite this, the AI’s preventive scaffolding and built-in instructional supports acted as a partial error filter, ensuring interface complexity did not fully compromise the final technical quality.

A critical finding emerged regarding engagement: a significant increase in data dispersion between pre-test and post-test. While one segment showed optimal adaptation, another recorded decreased commitment. These findings indicate an inconsistent user experience, suggesting that individual variables can prevent a generalized positive response despite the technical success achieved.

In conclusion, AI integration optimizes food design by reducing experimental margins of error and accelerating the innovation cycle. However, the gap between algorithmic power and user interface remains a challenge, highlighting that technical efficacy does not automatically guarantee a uniform learning experience. To ensure the integral adoption of these technologies, it is imperative to strengthen prior training and pedagogical onboarding. Future interventions should focus on enhancing student digital literacy and providing robust instructional support to mitigate the cognitive friction caused by complex interfaces, thereby aligning technical success with a high-quality educational experience.

Keywords:

Higher Education, Educational Innovation, AI-powered food design, Competency Assessment, Enhancing Learning.