The termination of more than 1,400 humanities grants during the early months of Donald Trump’s second presidential administration has come under intense examination following newly released deposition testimony from two staff members involved in the review process. Justin Fox and Nathan Cavanaugh, who worked with the Department of Government Efficiency initiative, acknowledged that artificial intelligence tools and keyword searches played a role in identifying grants potentially linked to diversity, equity, and inclusion.
Their testimony, presented as part of an ongoing lawsuit filed by several major academic organizations, has raised significant questions about how the federal government evaluated humanities projects and determined which ones would lose funding. According to court filings, the grants affected belonged to the National Endowment for the Humanities and represented roughly 97 percent of the agency’s active awards at the time. The combined value of the cancelled projects exceeded one hundred million dollars, making the decision one of the most sweeping funding reductions in the agency’s history.
The process behind those cuts has now become central to the legal challenge brought by scholarly groups who argue that the government acted improperly when it terminated funding for projects focused on topics such as race, gender, and LGBTQ communities. Deposition videos and transcripts reveal that Fox and Cavanaugh reviewed large spreadsheets listing hundreds of projects that had been approved during the previous administration.
Their task was to identify grants that might conflict with executive orders aimed at eliminating programs connected to diversity, equity, and inclusion initiatives within the federal government. However, the testimony also revealed that the staffers had difficulty clearly explaining what the term “DEI” actually meant while they were reviewing the projects. The combination of uncertain definitions and reliance on automated tools has intensified debate about whether the review process was rigorous, informed, or appropriate for evaluating academic work.
The AI-Driven Review of Humanities Grants
During his deposition, Justin Fox described how the review process frequently relied on digital searches and artificial intelligence tools to screen grant descriptions. Working with spreadsheets containing hundreds of project summaries, Fox said he searched federal grant databases using keywords he believed might indicate connections to diversity, equity, and inclusion topics. Among the words he used were “gay,” “BIPOC,” “indigenous,” “tribal,” “melting pot,” and “equality.” Any project that appeared to contain these terms was flagged for further examination.
Fox testified that he also used ChatGPT to analyze project descriptions. In some cases, he copied the summary of a grant proposal into the system and asked a simple question: whether the project related at all to diversity, equity, and inclusion. The answer produced by the AI system could influence whether a project was considered potentially incompatible with executive orders targeting DEI programs.
One of the most notable aspects of Fox’s testimony was his acknowledgment that he did not define the term “DEI” when presenting the question to the AI system. When attorneys asked how the system would know what qualified as diversity, equity, and inclusion without a definition, Fox admitted that he had not considered how the model interpreted the concept. His approach relied largely on the assumption that the system could identify relevant themes based on the wording of the project description.
Fox also explained how certain projects were classified as DEI-related during the review. In one example discussed during the deposition, a documentary project focusing on female Holocaust survivors was flagged because it highlighted gender perspectives and centered on the experiences of Jewish women. Fox described it as amplifying marginalized voices within a cultural group and said that the gender-based narrative contributed to its classification.
Another example involved a project examining historical events connected to the civil rights movement and Black communities. Fox indicated that it was identified as DEI-related because it focused on the experiences of a specific racial group. When attorneys questioned that reasoning, he clarified that the project related to diversity, equity, and inclusion themes because it centered on racial history.
WATCH: “How do you interpret DEI? Is any documentary about women DEI?”
— The Tennessee Holler (@TheTNHoller) March 11, 2026
Elon’s DOGE guy Justin Fox can’t explain what DEI is — or why they canceled a grant for a documentary about women during the Holocaust.🤔
Full: https://t.co/iWajV6Qjro pic.twitter.com/GeBZj5jm1b
Fox also noted that research involving LGBTQ topics frequently stood out during the review process. Grants referencing gender identity, gender fluidity, or LGBTQ history were often flagged because they appeared to align with what reviewers considered DEI subject matter. According to Fox, projects focusing on LGBTQ communities were seen as related to underrepresented groups, which made them more likely to attract scrutiny during the review.
The reliance on keyword searches and artificial intelligence became one of the most controversial elements of the process described in the depositions. Critics argue that humanities scholarship cannot be evaluated through simple keyword detection, particularly when topics such as history, culture, and identity are often intertwined with broader academic inquiry. The deposition testimony suggests that the screening process may have been more automated and less nuanced than typically expected in academic grant evaluation.
Defining DEI Proved Difficult for the Reviewers
While Fox described how projects were flagged for potential termination, his deposition also revealed uncertainty about the definition of diversity, equity, and inclusion itself. When attorneys asked him to explain what DEI meant in practical terms, Fox struggled to provide a clear answer. Instead, he referred broadly to the language contained in an executive order aimed at eliminating what it described as radical or wasteful DEI programs.
Fox acknowledged that he did not remember the specific details of that order when asked to elaborate during the deposition. His explanation relied mainly on the idea that DEI involved programs or initiatives associated with marginalized groups or discussions about social inequality. However, when pressed to provide a more precise definition, he said he did not have the exact language in mind.
Read : Elon Musk to Step Back from DOGE After Tesla Net Income Fell 71 Percent in the First Quarter
The difficulty in defining DEI became a central issue in the questioning by attorneys representing the academic organizations challenging the grant cancellations. Their argument suggests that if the individuals responsible for identifying DEI-related grants could not clearly explain the concept, it raises concerns about the criteria used to terminate funding.
Nathan Cavanaugh’s deposition echoed similar themes. As a political appointee working with the Department of Government Efficiency team through the General Services Administration, Cavanaugh said he participated in reviewing grants despite having no formal background in the humanities or academic research. Like Fox, he relied primarily on reading summaries of grant proposals and looking for language that appeared connected to diversity or identity topics.
Cavanaugh acknowledged that a discussion series examining LGBTQ experiences in military service was flagged during the review. When asked why, he said the reason was straightforward: the description explicitly referenced LGBTQ topics. Another grant examining the legacy of HIV and AIDS activism and its connections to feminist and queer perspectives was also flagged because it referenced gender and LGBTQ themes.
During the deposition, Cavanaugh explained that he believed it was possible to make judgments about such topics without specialized academic training. He suggested that a person could develop sufficient understanding through general reading and being informed about social issues outside traditional academic settings. However, when attorneys asked which specific books had influenced his thinking about these issues, Cavanaugh acknowledged that there were none.
This exchange became one of the more widely discussed moments from the deposition because it highlighted the informal nature of the criteria used to assess complex humanities projects. Critics argue that evaluating academic research typically requires familiarity with scholarly standards, peer review practices, and subject expertise. The admission that the reviewers lacked such backgrounds has strengthened arguments that the process may not have met those standards.
Legal Challenge and Academic Community Response
The deposition testimony surfaced as part of a lawsuit filed by several major scholarly organizations representing researchers in the humanities and social sciences. These groups argue that the government unlawfully terminated grants that had already been approved through the National Endowment for the Humanities’ established peer review process. According to the lawsuit, the cancellations disproportionately affected projects addressing topics related to race, gender, sexuality, and social justice.
Read : Federal Employee Says Replying To DOGE Mail is Like Living In North Korea
The National Endowment for the Humanities traditionally relies on panels of scholars and experts to evaluate grant proposals based on academic merit, research value, and public benefit. The organizations bringing the lawsuit argue that the decision to cancel hundreds of projects bypassed that system and replaced it with a politically driven review that did not involve subject matter experts.
Court filings also highlight procedural issues related to how the termination notices were sent. According to the documents, Fox sent many of the notifications using an unofficial Microsoft email account rather than the agency’s standard grant management system. The notices reportedly used nearly identical language and did not provide individualized explanations for why each project had been cancelled.
For scholars and cultural institutions that had already begun work on the grants, the cancellations created significant disruption. Many projects had secured partnerships with museums, libraries, and educational organizations, while others had already begun research, public programming, or documentary production. The sudden loss of funding forced some institutions to suspend work or seek alternative sources of financial support.
Within the academic community, the controversy has sparked broader debate about how humanities research is evaluated and the role of political priorities in shaping federal funding decisions. Supporters of the grant cancellations argue that the government has the authority to redirect funding away from programs it believes conflict with policy goals. Critics counter that humanities scholarship often explores topics related to identity, culture, and inequality, and that removing funding based on such themes risks undermining academic freedom and historical research.
The deposition testimony has intensified these discussions by revealing the mechanics of the review process in greater detail. The use of keyword searches, the involvement of reviewers without academic expertise, and the reliance on artificial intelligence tools have all become central points of controversy in the legal challenge.
As the lawsuit moves forward, courts will likely examine whether the grant cancellations complied with federal law governing agency decision-making and whether the process respected the established procedures used by the National Endowment for the Humanities. The testimony from Justin Fox and Nathan Cavanaugh has already become a key element in that evaluation, offering a rare look inside the decision-making process behind one of the most extensive funding reductions ever imposed on federal humanities programs.