Leveraging artificial intelligence and short educational videos to expand accessible cancer education for diverse communities

Background: Colorectal cancer (CRC) is the second leading cause of cancer death in the United States, yet many adults lack knowledge about screening, symptoms, and genetic risk. Traditional materials often do not reach diverse socioeconomic and cultural groups because of low health literacy and differences in media use. Many Americans, especially younger and lower income adults, spend several hours each day on screens. This creates an opportunity for mobile friendly and visually engaging education. Artificial intelligence (AI) offers a rapid way to create accessible and adaptable cancer information that fits current media habits. Methods: We created an early stage framework that uses AI to produce scalable and culturally adaptable cancer education. The approach uses generative video platforms, AI assisted scriptwriting, simplified language, and multilingual narration to create short films on prevention, screening, and genetic testing. The framework aims to increase knowledge and early detection behaviors, reduce production time and cost, support mobile and screen based viewing, and adapt content across languages, cultures, and socioeconomic settings. Initial prototypes include videos on lifestyle risk reduction, family history, genetic testing, and timely screening. Results : Our team produced AI assisted scripts for a storyboard, a storybook concept, and a film to educate African American (AA) communities about CRC. Four focus groups with a total of 27 AA and African immigrant adults were conducted in person and on Zoom through the AltaMed patient network. Participants reported that the script was clear, easy to understand, and communicated the main ideas. Community specific statistics were motivating and relevant. Participants asked for more information about the screening experience and reassurance to reduce fear. Their feedback informed the development of a script board with key message cards that address barriers and essential CRC information for AA audiences. Conclusions: AI assisted workflows can support the creation of culturally relevant CRC education materials while maintaining clarity and community alignment. This approach shows strong potential for producing high quality materials for populations underserved by traditional communication. The next phase will assess feasibility, acceptability, comprehension, and intent to seek CRC screening among diverse audiences, with planned use in community outreach, patient navigation, and digital education programs.