Elon Musk's Department of Government Efficiency (DOGE) operatives relied on ChatGPT and their own uninformed judgments to cut more than 1,400 grants from the National Endowment for the Humanities, according to newly released video testimony. The depositions, from a lawsuit by academic associations, reveal that two young DOGE officials with no government or grant administration experience made sweeping funding decisions.
Justin Fox and Nathan Cavanaugh, the DOGE officials in question, used a ChatGPT prompt asking whether each grant related to DEI (diversity, equity and inclusion). They instructed the AI to respond factually in under 120 characters, beginning with 'Yes.' or 'No.' Fox admitted during his deposition that he considered a documentary about Black civil rights as violating Trump's executive order on DEI because it 'focused on a singular race.' He struggled to define DEI, basing his understanding solely on Trump's executive order.
Cavanaugh, 28, co-founder of Brainbase and Special, admitted he had no experience in scholarly or peer review processes. He claimed a person could make judgment calls from being 'well-informed' but could not name any specific books consulted. The only grants spared were those related to America's 250th anniversary and the 'National Garden of Heroes' – both Trump priorities.
The Modern Language Association's executive director, Paula M. Krebs, described DOGE's actions as 'haphazard and unlawful,' carried out by 'unqualified agents' who 'undermined the separation of powers.' The lawsuit alleges that the cuts, which eliminated tens of millions in public funding within a month, were based on flawed AI analysis and personal bias rather than expertise.
