In the face of accelerating language loss, can artificial intelligence help preserve linguistic diversity? Researchers are exploring an innovative application of Large Language Models (LLMs) to automatically generate “glosses,” which are essentially linguistic annotations, for endangered languages. These annotations are like detailed footnotes for each word, explaining its meaning and grammatical function. Traditionally, creating these glosses is a labor-intensive process requiring deep linguistic expertise. Automating this process with AI has the potential to significantly speed up language documentation efforts. The research tested different methods of prompting LLMs to generate accurate glosses, finding that providing the models with relevant examples improves their performance. Specifically, the most effective technique involved showing examples that shared character sequences with the target sentence, hinting at shared morphological structures. The results suggest that AI models are increasingly able to discern patterns even in languages they haven’t explicitly been trained on. While AI isn't a perfect solution, it offers a promising new tool in the fight to save the world's linguistic heritage, potentially making language documentation more efficient and scalable. These AI glossers, while not a replacement for human linguists, can be valuable assistants, especially in extremely low-resource settings. This technology could help preserve not just words, but the cultural knowledge and traditions embedded within these languages.