Teaching AI for Social Good: Youth-Centered Critical AI Literacy

This study examines the design and implementation of the D3 (Decode, Discover, Design) curriculum, a culturally relevant AI and data literacy program for middle school students. Sixty students from a low-SES, predominantly Hispanic school in the southern United States participated in this project. The curriculum integrated three complementary perspectives: AI literacy, AI for social good, and culturally relevant computing. During the Decode and Discover phases, students learned foundational AI and NLP concepts while analyzing real-world community data. In the Design phase, students created digital artifacts that addressed issues. Mixed methods results revealed significant gains in AI knowledge and more positive attitudes toward AI. Students also demonstrated agency, civic engagement, and awareness of algorithmic bias. The study demonstrates the potential of culturally relevant computing and AI for social good, offering a transferable model for community-anchored, equity-focused AI education.