This contribution presents the design of an AI-supported learning activity developed to support conceptual revision in high school or undergraduate chemistry courses. The activity is based on structured student interaction with formal chemical representations and guided questioning, aiming to promote active recall, conceptual precision, and low-stakes self-assessment. Rather than relying on direct interaction with autonomous AI agents, the activity is implemented through an application programming interface (API) architecture, in which pedagogical logic and task progression are explicitly controlled at the instructional design level. This design choice allows the instructor to constrain AI behavior, reduce unintended variability in responses, and ensure alignment with specific learning objectives. From a technical perspective, the API-based approach enhances reproducibility, scalability, and sustainability, facilitating integration into web applications or learning management systems and enabling flexible model substitution without altering the instructional structure. From the students’ perspective, the activity functions primarily as an active revision tool, encouraging engagement with key chemical concepts through iterative questioning and immediate feedback. Preliminary classroom use suggests that students perceive the activity as supportive for consolidating terminology, recognizing structural features, and checking their understanding outside scheduled teaching time. The poster discusses the instructional design principles underlying the activity, the pedagogical advantages of an API-based approach compared to direct agent interaction, and the potential transferability of the design to other chemistry topics and educational contexts.
Lombardo, R., Maggio, A. (2026). From AI Agents to APIs: Pedagogical Design and Educational Advantages of an API-Based Learning Activity in Chemistry. In ICCE ECRICE 2026 (pp. 154-154). Turkish Chemical Society.
From AI Agents to APIs: Pedagogical Design and Educational Advantages of an API-Based Learning Activity in Chemistry
LOmbardo, Renato
;Maggio, Antonella
2026-01-01
Abstract
This contribution presents the design of an AI-supported learning activity developed to support conceptual revision in high school or undergraduate chemistry courses. The activity is based on structured student interaction with formal chemical representations and guided questioning, aiming to promote active recall, conceptual precision, and low-stakes self-assessment. Rather than relying on direct interaction with autonomous AI agents, the activity is implemented through an application programming interface (API) architecture, in which pedagogical logic and task progression are explicitly controlled at the instructional design level. This design choice allows the instructor to constrain AI behavior, reduce unintended variability in responses, and ensure alignment with specific learning objectives. From a technical perspective, the API-based approach enhances reproducibility, scalability, and sustainability, facilitating integration into web applications or learning management systems and enabling flexible model substitution without altering the instructional structure. From the students’ perspective, the activity functions primarily as an active revision tool, encouraging engagement with key chemical concepts through iterative questioning and immediate feedback. Preliminary classroom use suggests that students perceive the activity as supportive for consolidating terminology, recognizing structural features, and checking their understanding outside scheduled teaching time. The poster discusses the instructional design principles underlying the activity, the pedagogical advantages of an API-based approach compared to direct agent interaction, and the potential transferability of the design to other chemistry topics and educational contexts.| File | Dimensione | Formato | |
|---|---|---|---|
|
P069.pdf
Solo gestori archvio
Tipologia:
Versione Editoriale
Dimensione
393.43 kB
Formato
Adobe PDF
|
393.43 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


