HathiTrust lands $5M Mellon grant to open 19M digital books to AI research
What's the deal? HathiTrustDealroom has a profile for this one. Try Dealroom → has received US$3.24M, five-year grant from the Mellon Foundation to build AI-enabled services that expand access to its collection of 19 million digitised volumes. The initiative, called Transparent Books: AI Infrastructure for Discovery and the Public Interest, will run alongside Northeastern University LibraryDealroom has a profile for this one. Try Dealroom → and Authors AllianceDealroom has a profile for this one. Try Dealroom →, a nonprofit that advocates for authors making their work publicly accessible.
What's the endgame? The project aims to give noncommercial researchers lawful access to in-copyright book data without relying on corporate mediation. Planned work includes using AI to improve metadata, text quality, and search, plus providing verified resources for text mining and language model development.
Why now? The effort builds on the 2024–2026 Public Interest Corpus project, led by Authors Alliance and Northeastern, which explored lawful ways for libraries to open book data to computational research. Authors Alliance executive director Dave Hansen said its role is to help set principles for corpus use and to convene stakeholders "so their perspectives inform project design."
What could go wrong? Applying AI to in-copyright collections raises legal and attribution questions. The partners say the goal is to lower barriers for academic researchers while ensuring authors' works are properly attributed and accessible.
The signal: At $5 million, the grant ranks among the larger public-interest research awards in its category — above the 90th percentile for US grant rounds in the field. It reflects a broader push to make libraries, rather than corporate platforms, the gatekeepers of lawful data for AI training and scholarly discovery.
Read more: knowledgespeak.com
Image credit: Open.Michigan
