Piotr Adamczyk - Working with cultural heritage collections and scholarly aggregations: Will we ever have enough metadata?

When: 2024-02-26 14:00-16:00 (Tallinn time)
Where: online

The event is public via zoom: https://zoom.us/j/94927755198?pwd=VWpWU25DY1JlVFJsbDF4aUtyL3Zvdz09

Abstract – We can build magical experiences with cultural heritage data - VR walks through a Van Gogh bedroom, AI generating countless images in any artistic style, collections of “all” of the works by “any” artist. The simple narratives in these projects reach the widest audience, but can we be more honest about the limitations without ruining the magic? Missing records from the British Museum, lack of provenance for Native American collections, no transparency in image training sets; Our tools let us satisfy our ambition for magic without having to be honest about what’s behind the curtain, synthetic metadata and hallucinating LLMs. Exposing how the trick is done won’t make these things any less powerful and may help us focus on the things that will make them even better.

Additional information – Piotr Adamczyk is a Computer Scientist and Librarian. Starting with an analyst position at The Metropolitan Museum of Art, as a Product Manager at the Google Cultural Institute, and Director at Artstor/JSTOR, Piotr has focused on how open/linked collections information can broaden the reach of cultural heritage institutions. Now at DataCite, Piotr is focused on how datasets move through scholarly infrastructure and how to make sure they are used and preserved.