How can we explore pressing issues through lived experiences of average people? Our climate change work has had real impact by engaging audiences with people just like them. To deliver this impact, we combine crowdsourced video with data visualisation. Now, we are developing a new technology, cross-cultural AI. It automatically translates & analyses user-generated video to help generate culturally-specific insights. As filmmakers, we increasingly draw on crowdsourcing to generate content, ethnography to relate these hints and glimpses of life to issues & indexes that matter, and data visualisation to draw it all together -- while keeping people's stories at centre. AI can recognise objects, people, actions; culturally-specific datasets plus crowdsourcing mean that creative practitioners could work more easily with participants globally. While our work focuses on climate change, the AI-driven platform we're developing could be applied to ranges of content, contexts, projects & products. Our approach combines quality of insights derived from user-generated content, with quantity coming with scale: Our Nine Earths project ([http://nine-earths.net][0]) drew from hundreds of hours of video from 12 countries. AI can now help translate languages, identify relevant content & recognise patterns to say something meaningful, especially when combined with participants' input and verification. [0]: http://nine-earths.net