Memory without Origin: Is AI Stealing Our History?

Описание: Tech giants are now approaching research libraries to demand their most valuable "training data," often targeting undigitized and unprocessed holdings. But there is a massive risk: once historical data is absorbed into a "General Model," it is often impossible to remove, meaning institutions lose control over their history forever.

In this video, we break down the new UVA Archival AI Protocol, which designed to protect cultural heritage in the age of AI.

Developed by Leo S. Lo (Dean of Libraries, University of Virginia), these documents argue that digitization funding and AI training rights must be treated as separate decisions to prevent short-term financial relief from locking in a long-term loss of rights.

We explain the critical difference between "Retrieval-Based" access (which is often safe) and "General Model Training" (which absorbs content irreversibly).

Learn why libraries are now demanding "provenance"—the ability to trace an output back to a specific item—and the "right to stop" before handing over their collections.

Resources & Links:
• UVA Archival AI Protocol: https://doi.org/10.18130/5dqf-9w86
• Author: Leo S. Lo, University of Virginia Library.   / leoslo