Particle.news
Download on the App Store

Depositions Reveal ChatGPT-Driven Review That Helped Cancel 1,400+ NEH Grants

New depositions describe AI screening by nonexperts, a process plaintiffs say bypassed NEH peer review.

Overview

  • Testimony from Nathan Cavanaugh says projects were flagged for cancellation when descriptions included terms like LGBTQ, including a veterans discussion series flagged solely for using the word.
  • Justin Fox said he ran grant summaries through ChatGPT using a prompt that forced yes/no DEI labels with brief explanations after keyword scans for terms such as Black, homosexual, and LGBTQ+.
  • Both Fox and Cavanaugh acknowledged they lacked humanities backgrounds, did not consult NEH peer reviewers, and struggled to define DEI beyond citing the administration’s executive order.
  • Court filings report that more than 1,400 active NEH grants worth over $100 million were terminated, eliminating roughly 97 percent of the agency’s active awards.
  • The Modern Language Association, American Council of Learned Societies, and American Historical Association are suing, and filings allege some termination notices were sent from an unofficial email with boilerplate language.