Classroom-Ready AI in the Humanities: An Ethics-by-Design e-Course and a STE(A)M Digital Lab for Teacher Professional Development

Abstract

European policy expects AI-supported, data-informed teaching, yet distance PD often treats AI literacy, pedagogy, and ethics as parallel tracks. This study presents TENSILE, an Erasmus+ KA2 model for ethics-by-design, humanities-anchored AI PD that integrates the Community of Inquiry (CoI), TPACK, and evidence-based PD levers. The design comprises a six-module asynchronous Moodle e-course and a STEAM Digital Lab with curated tools, templates, and 25 classroom-ready use cases in language and literature-from rule-based OULIPO generators to lightweight Python/NLP and MLforKids to Scratch classifiers that keep model behaviour observable. Safeguards (privacy, transparency/provenance, fairness, and human oversight) are embedded as routine gates, with public artefacts (model/data cards, decision logs, human-vs-model tables) supporting critique and transfer. The theory of change aligns with EU risk-based AI governance and international guidance, focusing on enhancing AI literacy, integrating techno-pedagogical approaches, and promoting ethical practice. Evaluation triangulates pilot artefacts with validated self-report instruments (TAICS, AICO_edu) to examine competence shifts and fidelity of enactment. The contribution is a replicable, low-bandwidth-friendly template that elevates the humanities as first-class sites for AI-mediated inquiry and supports scalable mentoring and community building across diverse infrastructure contexts.