Elon Musk's DOGE Officials Relied on ChatGPT for Sweeping Grant Cuts
Depositions from Elon Musk's Department of Government Efficiency team have revealed a startling reliance on artificial intelligence and largely uninformed judgments to make sweeping decisions about federal grant funding. More than ten hours of newly released video testimony from January uncovers how two young DOGE operators abruptly choked off more than 1,400 grants at the National Endowment for the Humanities, eliminating tens of millions of dollars in public funding within less than a month.
ChatGPT Became the Arbitrator of Grant Decisions
During his hours-long deposition, DOGE official Justin Fox admitted to using ChatGPT as an intermediary step before slashing grants. The team used a specific prompt: "Does the following relate at all to DEI?" followed by instructions to "Respond factually in less than 120 characters" and begin with 'Yes.' or 'No.' with a brief explanation. This prompt, included in discovery materials along with emails, spreadsheets and text messages, became the primary filter for determining which grants would survive.
Fox, who previously worked at a private equity firm before joining DOGE, demonstrated concerning gaps in understanding during questioning. When reviewing a grant for a documentary about Black civil rights, he agreed the project violated Donald Trump's executive order on DEI because it "focused on a singular race." He stated: "It is not for the benefit of humankind. It is focused on this specific group, or a specific race, here being Black."
Officials Struggled to Define Basic Concepts
Perhaps most revealing was Fox's inability to articulate what DEI actually means. In a painful, minutes-long exchange during his deposition, full of pregnant pauses and heavy sighs, Fox said his "understanding" of DEI was based on Trump's executive order targeting diversity, equity and inclusion efforts across government. But when pressed, he admitted: "I can't remember" what was in it.
Fox's colleague Nathan Cavanaugh, 28, co-founder of Brainbase and later a business partner with Fox, admitted he had no experience in scholarly or peer review processes. When asked which books informed his judgments about grants, Cavanaugh responded bluntly: "There were no books."
"Craziest" Grants Targeted LGBTQ+ Projects
Fox compiled what he considered the "craziest" and "other bad" grants using three dozen keywords including "LGBTQ," "BIPOC," "Tribal," "ethnicity," "gender," "equality," "immigration," "citizenship" and "melting pot." More than two dozen grants deemed the "craziest" were related to LGBT+ projects.
In his deposition, Fox acknowledged the list reflected his "subjective" judgment, stating: "'Crazy' is one way of saying it. 'Most incriminating' is another way." One grant Cavanaugh reviewed concerned a book exploring the legacy of HIV and AIDS activism and prison abolition, which he flagged because it "references feminist and queer insights into prison abolition and LGBTQ studies."
Six-Figure Salaries and No Remorse
Both officials received substantial compensation for their work, with Fox earning $150,000 and Cavanaugh receiving $120,000. When asked if he felt any remorse for the grantees impacted by their decisions, Fox responded: "Sorry for those impacted, but there is a bigger problem, and that's ultimately — the more important piece is reducing the government spend."
Cavanaugh was even more direct, cutting off a question about regret before it was finished with a simple: "No." He added: "I think it was more important to reduce the federal deficit from $2 trillion to close to zero," while admitting they didn't come close to achieving that goal.
Legal Challenges and Broader Implications
The depositions stem from a lawsuit filed by the Modern Language Association, the American Council of Learned Societies and the American Historical Association. According to Modern Language Association executive director Paula M. Krebs, evidence in the case exposes DOGE's "haphazard and unlawful actions" from "unqualified agents" who "undermined the separation of powers and denied the American people access to vital public programming and research."
Despite DOGE making 29,000 cuts to government last year, federal spending didn't decrease under their watch. The bulk of DOGE's work — including devastating cuts to foreign aid recipients — ultimately amounted to very little within the scope of overall government spending, raising serious questions about the initiative's effectiveness and methodology.
