DANS is exploring AI-powered suggestion of keywords, taken from controlled vocabularies, based on the free-text description provided by the depositors of datasets in the Data Station Social Sciences and Humanities. This is part of a broader trend, as is also apparent in LTER-LIFE, where datasets archived in the Data Station Life Sciences are also automatically enriched with AI-generated metadata. This session will showcase both features and frame them in a discussion on the future of metadata against the background of developments in AI.