They walked into the National Endowment for the Humanities in March 2025. One was a college dropout who idolized Peter Thiel. The other was a private equity associate with a commerce degree and a German minor. Neither had ever administered a grant. Neither had read a humanities proposal. Neither knew what the Administrative Procedures Act was. They were just two guys from Elon Musk’s world armed with ChatGPT and six weeks to gut a federal agency. By April 1st, $100 million in grants was gone. Sixty-five percent of the staff was fired. The agency’s own employees didn’t even know it was happening. One NEH staffer emailed a grantee: “I’m terribly sorry to tell you that DOGE did indeed cancel your award. NEH staff, like myself, didn’t realize it was happening.”
Their method was as crude as it was discriminatory. Justin Fox created a keyword checklist and fed it into ChatGPT to screen every grant. The list read like a target sheet: LGBTQ, homosexual, tribal, immigrants, diversity, inclusion, queer, transgender, non-binary, indigenous, native, diaspora, refugees, asylum, equity, marginalized, BIPOC, social justice — and just in case they missed anything, the word “gay” at the end. Notably absent: white, Caucasian, heterosexual. When a lawyer asked Fox why “heterosexual” wasn’t on the list, he shrugged: “Very well could have put heterosexual.” He never did. When asked if he did anything to ensure ChatGPT wouldn’t discriminate on the basis of sex or religion, his answer was chilling: “We did not need to do that.”
PAID SUBSCRIBERS CAN ACCESS A DEEPER DIVE BELOW
