The college of william and mary in collaboration with old dominion university and the internet archive, will identify and preserve web resources that are referenced in archived US local and national…

The Institute of Museum and Library Services awarded $1.5 million to the College of William and Mary on September 1, 2026, under the National Leadership Grants program (CFDA 45.312) to identify and preserve web resources referenced in archived US local and national television news.

The College of William and Mary, in collaboration with Old Dominion University and the Internet Archive, will develop novel artificial intelligence tools to detect web citations in US television news broadcasts and build a public database linking archived web pages to the TV programs that cite them. Core deliverables include multi-modal machine learning and deep learning models for web citation detection and context classification; a large-scale web-to-TV news citation database; context banners on archived web pages referenced in TV news; and analysis of information flow patterns from the web to broadcast media. The project will enable journalists, researchers, and the public to trace how information moves between online and broadcast media.

Performance occurs in Williamsburg, Virginia, with a period of performance extending through August 31, 2029. The award is structured as a Project Grant, the standard assistance type under National Leadership Grants, which supports projects generating models, tools, research findings, or practices that can be widely adapted, scaled, or replicated across the library and archival professions.

The college of william and mary in collaboration with old dominion university and the internet archive, will identify and preserve web resources that are referenced in archived US local and national television news. This project will provide context for both collections and enable the public, journalists, and researchers to trace how information moves between the web and broadcast media. The main research outcomes are developing novel artificial intelligence tools for effectively identifying web resources that are referenced in US TV news and building a public database for them, adding context banners on archived pages that are referenced in TV news, and investigating the web to TV news information flow. The project's tasks include developing novel multi-modal machine learning/deep learning web citation detection and context classification models; developing a large-scale web/TV news citation database; and linking archived web pages to citing TV news programs; characterizing information flowing from the web to US TV news.