The increasing development of Artificial Intelligence demands a comprehensive and interdisciplinary approach to assess its ethical, legal, relational, and social impact. However, bridging the epistemic gap between engineering and social sciences remains a challenge. This project introduces the Challenge-Device-Synthesis (RDS) methodology within the Bachelor's Degree in Artificial Intelligence at the Universitat Autònoma de Barcelona. RDS fosters knowledge hybridization by engaging students in real-world challenges, where they develop AI-based devices that serve as objects of study for assessing their societal implications.
The initiative initiated consists of two parts: 1) The design of the methodological body (already carried out during Q3-Q4 2024, and Q1 2025), which is the core of this contribution, and 2) the validation of the proposed design, including the deployment of the devices in the context of real-life settings to resolve the proposed challenges.

In the proposed approach, students’ tasks are distributed in two 6-ECTS subjects during the 2nd semester of their 3rd year, namely: 1) Project Synthesis (3-ECTS), in which a collection of curated challenges from private companies are carried out by teams in a software development project -with a github as an outcome-, and 2) Social Innovation (3-ECTS), in which the students aim at designing an experimental space for their devices in a real-life setting, which will serve as the basis for the discussion of the dimensions of social impact -with a 10-page essay as an outcome-.

A key element is the integration of interdisciplinary assessment group, comprising experts from 4 different disciplines: psychology, law, philosophy and computer science. This group has the mission to conduct two structured evaluation sessions -mid-term at week 7th and final assessment at week 15th-, applying a multidimensional framework to analyze the AI prototypes. The committee is commissioned to examine the devices from the technical, ethical, legal, relational, and social impact perspectives. The assessment process considers compliance with existing regulations (including the new AI Act), ethical concerns related to fairness and transparency, the impact on human relationships and interactions, and broader societal consequences.

The proposed framework has provided as a result the restructuring of the 6th semester of the Degree of AI in an integrated challenge-based approach, connecting 12-ETCS from an inter-disciplinary perspective. It attracted 4 interested industrial partners, and particularly Cruïlla Music Festival Barcelona (event be yearly visited by more than 200,000 attendees during the first week of July), providing two specific industrial challenges for experience personalization using generative AI, and for community engagement through sentiment analysis in the Internet. The real-life setting appears essential for the substantiation of the challenges through the design and implementation of a set of functional AI prototype systems, ensuring their applicability and relevance. The results of the interaction of the attendees with the prototypes during the Cruïlla Music Festival will be reported in a future contribution.

As main conclusion, the proposed methodological design may endow AI Higher Education Programs with a transversal approach, focusing both on the high-quality software development tasks while addressing fundamental social impact dimensions in a solid curricular integration.

Keywords: Artificial intelligence, Challenge-Based Learning, Ethics, Transdisciplinarity.