Study published by the National Library of the Netherlands (KB) and the Europeana Foundation: “This document identifies a pathway to establishing the core of the European Books Data Commons: a shared infrastructure that would make the full text of millions of digitised public domain books held by libraries across Europe available for re-use in forms optimised for AI developers and researchers working with large-scale language datasets. If implemented, the EBDC would constitute a significant contribution to the European Commission’s 2025 Data Union Strategy, which aims to ensure that European AI developers have access to high quality data including cultural heritage collections.
The remainder of the document is structured as follows. Section 2 sets out the proposition — the demand for digitised public domain books and the supply-side constraints that currently prevent access to existing collections. Section 3 presents the recommended implementation scenario, arrived at through a process of developing, testing, and progressively refining a set of options with library partners and the steering group. Section 4 addresses the governance of the system and its relationship to related European initiatives. Section 5 outlines the way forward, including a two-track implementation approach, an integrated timeline, and indicative cost estimates. Section 6 makes the case for acting now to build the European Books Data Commons…(More)”.
