Judge Blocks Mass Grant Termination After Shocking Discovery

Person wearing a black shirt with 'Doge' printed on it, standing outdoors at night

A federal judge has struck down the Trump administration’s mass termination of thousands of humanities grants, ruling that Elon Musk’s Department of Government Efficiency discriminated based on race and sex using artificial intelligence and keyword searches.

Story Snapshot

Judge Colleen McMahon ruled DOGE “blatantly used” race and sex to terminate congressionally approved National Endowment for the Humanities grants in violation of the Constitution
Inexperienced DOGE staffers in their 20s used ChatGPT and keyword searches for terms like “DEI,” “LGBTQ,” and “equity” to justify cutting thousands of grants without reviewing applications
Holocaust and Jewish women’s history projects were flagged and terminated using the discriminatory process, raising concerns amid rising antisemitism
The ruling blocks enforcement of the terminations and potentially restores millions in funding to scholarly organizations including the American Council of Learned Societies and American Historical Association

Judicial Rebuke of Executive Overreach

U.S. District Judge Colleen McMahon issued a scathing 143-page ruling declaring the Department of Government Efficiency’s termination of National Endowment for the Humanities grants unconstitutional and unlawful. The judge determined that DOGE violated the First and Fifth Amendments by discriminating based on protected characteristics including race, sex, and ethnicity. McMahon found the process did not conform to NEH’s ordinary peer-review procedures and exceeded executive authority by unilaterally blocking funds Congress had already appropriated, violating the 1974 Impoundment Control Act established after similar Nixon-era overreach.

AI and Keywords Replace Expert Review

Depositions of former DOGE staffers revealed a troubling process that replaced decades of scholarly peer review with automated keyword searches and artificial intelligence. Young staffers in their twenties with no humanities expertise used ChatGPT to generate rationales for terminating grants flagged by searches for terms like “DEI,” “BIPOC,” “equity,” and “LGBTQ.” Judge McMahon noted the “mere mention of a woman” or references to minority communities triggered terminations. This approach targeted projects documenting Black history, Native American experiences, Asian American studies, and even Holocaust research, raising alarms about both discrimination and the erosion of merit-based academic funding.

Constitutional Concerns Beyond Political Divisions

The ruling exposes a fundamental problem that transcends partisan politics: unelected bureaucrats wielding unchecked power over congressionally mandated programs. While conservatives rightly demand efficiency and oppose wasteful DEI initiatives, many will question whether junior staffers using AI to override Congress respects constitutional separation of powers. The Impoundment Control Act exists precisely because President Nixon’s attempts to unilaterally block appropriated funds violated legislative authority. Citizens across the political spectrum should be concerned when executive branch officials, however well-intentioned their efficiency goals, substitute algorithmic keyword scans for the transparent processes and statutory authority required by law.

Broader Implications for Government Reform

The decision weakens DOGE’s credibility as an efficiency reform body and may trigger congressional oversight of its operations across multiple agencies. The ruling signals clear limits on executive authority to terminate appropriated funds without following established legal procedures, even when pursuing legitimate cost-cutting objectives. Scholars and institutions affected by the terminations stand to regain millions in restored funding, though uncertainty remains about potential appeals to the Second Circuit Court. The case demonstrates how reform efforts, when implemented through shortcuts that bypass constitutional constraints and rely on inexperienced staff using untested AI methods, can backfire spectacularly and undermine the very efficiency goals they claim to serve.