Ownership
Who holds rights over language materials, the things made from them (dictionaries, recordings, models), and any value those produce. It is easy to lose track of in AI work, where data passes through many hands: a community may own its recordings yet find that a dataset or model built from them is controlled by someone else (see Data Scraping and Transkribus). Clear ownership is a precondition for data sovereignty and for a community to benefit from its own language work rather than having it extracted.