The integration of artificial intelligence (AI) technologies with traditional research methodologies presents significant opportunities for enhancing educational and social interventions while maintaining scientific rigor and community engagement. However, current approaches often lack systematic frameworks for ensuring community ownership, ethical implementation, and sustainable social impact. This study introduces and validates the Community-Based AI Development (CBAID) framework through comprehensive analysis of five diverse AI projects implemented during the Accadia Winter School initiative, focusing on methodological innovation, replicability, and social impact. We employed a multiple case study design analyzing five AI projects: G.A.M.E.S.-I.N. (health promotion), AI4Citizens (digital governance), LLM-Didattica (educational technology), DACSE (health communication), and AI-Enhanced Cybersecurity Training. Data collection included project documentation, stakeholder interviews (n=47), focus groups (n=8), surveys, and observational records. Cross-case analysis identified common patterns and framework validation evidence. All five projects demonstrated successful CBAID framework implementation with significant positive outcomes. Community engagement indicators showed high satisfaction (4.3/5.0) and meaningful participation in decision-making. Individual outcomes included enhanced knowledge, skills, and self-efficacy across domains. The framework showed strong transferability across diverse contexts with systematic adaptation guidance.
Toto et al. (Thu,) studied this question.