Artificial intelligence (AI) applications can profoundly affect society. Recently, there has been extensive interest in studying how scientists design AI systems for general tasks. However, it remains an open question asto whether the AI systems developed in this way can work as expected in different regional contexts while
simultaneously empowering local people. How can scientists co-create AI systems with local communitiesto address regional concerns? This article contributes new perspectives in this underexplored direction atthe intersection of data science, AI, citizen science, and human-computer interaction. Through case studies,
we discuss challenges in co-designing AI systems with local people, collecting and explaining communitydata using AI, and adapting AI systems to long-term social change. We also consolidate insights into bridgingAI research and citizen needs, including evaluating the social impact of AI, curating community datasets for
AI development, and building AI pipelines to explain data patterns to laypeople.@en