The article discusses a volunteer network of interpreters who are working to make the languages of refugees and other marginalized groups accessible to artificial intelligence (AI) systems. The key points are: 1) AI systems often struggle to understand languages spoken by refugees and other marginalized communities due to a lack of training data. 2) A global network of volunteer interpreters called Translators without Borders is working to address this issue by collecting and annotating data in these languages. 3) The volunteers transcribe and translate audio recordings, which are then used to train AI models to better understand these languages. 4) This effort aims to make AI technologies like speech recognition and machine translation more inclusive and accessible to marginalized communities. 5) The project has already collected data in over 200 languages and is helping to bridge the language gap in AI systems.