Abstract:

A recurrent challenge in AI education is the perceived gap between technical expertise and social reflection, with the latter often seen as less important. This reflects a broader divide in engineering education between “problem-solving” cultures and critical inquiry traditions. From a Science and Technology Studies perspective, this divide is analytically untenable. AI systems are not isolated artifacts; they are sociotechnical assemblages that interact with institutions, norms, and infrastructures. The challenge is to redesign learning environments where technical and social reasoning are co-constitutive.

This paper introduces the Techno-Values Committee as a pedagogical innovation that transforms assessment into a site of sociotechnical integration. Developed within the Challenge–Device–Synthesis framework, it was implemented in the sixth semester of the Bachelor's Degree in Artificial Intelligence at Universitat Autònoma de Barcelona. Two courses (Project Synthesis and Social Innovation) were merged into a single module where students developed AI prototypes while analyzing their social impact.

Over the semester, eight teams designed AI devices addressing real-world challenges. Each project required both a technical deliverable (repository, architecture, documentation) and an analytical essay examining ethical integrity, regulatory compliance, stakeholder implications, and social effects.

The Techno-Values Committee was the central moment of the process. Composed of experts in psychology, philosophy and law, the committee convened to evaluate near-final prototypes using a five-dimensional rubric: technical robustness, regulatory adequacy, ethical soundness, human–machine relational effects, and societal impact. Students had to justify architectural decisions in legal terms, explain model bias mitigation strategies in ethical terms, and articulate relational implications using psychosocial language. Assessment became a performative event, making technical artifacts publicly accountable.

Our analysis focuses on three aspects: first, the design of the committee, deliberation protocol, and coordination; second, the impact on student reasoning, observed in final reports and feedback, where students shifted from treating social analysis as an “external commentary” to recognizing it as integral to practice; and third, how such an interdisciplinary evaluative device could evolve into a stable curricular infrastructure.

The findings indicate that the committee format led to measurable epistemic shifts: students demonstrated increased anticipatory thinking about unintended consequences, greater sensitivity to regulatory risk, and a more relational understanding of AI deployment. The experience aligned with Responsible Research and Innovation principles, particularly anticipation, reflexivity, inclusion, and responsiveness. Students described the evaluative encounter as qualitatively different from traditional grading; more of an accountability exercise before a plural tribunal.

By conceptualizing assessment as a sociotechnical arena, the Techno-Values Committee repositions evaluation as a driver of epistemic integration. The model offers a transferable framework for AI and other engineering degrees seeking to bridge the gap between technical competence and social responsibility, showing that interdisciplinarity can be institutionalized not only in content but also in academic judgment.

Keywords:

Education, Technology, AI, Pedagogical methods, Innovations, Challenge based learning.