What could be meant by a metadata schema?

A metadata schema defines attributes (properties), their cardinality, the multiplicity between objects, and the relationships that connect them. It provides the formal structure necessary to describe resources consistently and to ensure semantic clarity across systems. Relationships, in particular, must be explicitly defined within the schema and carefully selected during the curation process to guarantee meaningful and interoperable connections.

Persistent identifiers (PIDs) for datasets, individuals, and organizations can also be understood as components of metadata schemas. They enable unique identification, reliable referencing, and the creation of robust links across distributed infrastructures. Furthermore, clearly specified exchange formats embedded in metadata schemas make it possible to present the same content consistently across different repositories and portals.

Against this backdrop, this poster revisits the foundational elements of metadata schemas, drawing on the DataCite schema as a practical example. At the same time, it raises a critical question: Do we still need all of this in the age of AI? As artificial intelligence increasingly derives context and complex relationships automatically, the role, function, and future relevance of formal metadata schemas invite renewed reflection and discussion.