Over the course of 2 years a global technology education nonprofit engaged ~ 20,000 under-resourced 3rd-8th grade students, parents and educators from 13 countries in a multi-week AI competition. Families worked together with the help of educators to identify meaningful problems in their communities and developed AI-prototypes to address them. Key findings included: (1) Identifying a high level of interest in underserved communities to develop and apply AI-literacy skills; (2) Determining curricular and program implementation elements that enable families to apply AI knowledge and skills to real problems; (3) Identifying effective methods of engaging industry mentors to support participants; (4) Measuring and identifying changes in self-efficacy and ability to apply AI-based tools to real-world problems; (5) Determining effective curricula around value-sensitive design and ethical innovation.