0
0
DOGE’s Deposition Chaos: How Musk’s ‘AI‑Powered’ Grant Cuts Became a Comedy of Errors – And Saved Nothing.
In the bizarre world of DOGE, the government’s Department of Government Efficiency, a 2026 DOGE deposition is shedding new light on one of the most chaotic federal grant review efforts in recent history. Testimony from DOGE staffers revealed how artificial intelligence tools like ChatGPT were used to identify and terminate federal grants — particularly projects tied to diversity, equity, and inclusion (DEI) and LGBTQ education — despite employees admitting they had little understanding of the programs they were cutting. Musk’s pet project to streamline government spending, critics say, ended up looking less like efficiency and more like what one observer described as “letting an intern use a wrecking ball.” The deposition testimony now offers the public a rare inside look at what happens when government officials attempt to wield AI without understanding what it is actually doing.
The DOGE deposition testimony revealed that roughly 1,400 federal grants were terminated using AI-assisted reviews — yet staff admitted the cuts saved no money and triggered legal challenges.
DOGE Investigation Series
This article is part of The Dystopeka Files investigation into the Department of Government Efficiency (DOGE), including the DOGE Social Security data scandal and related reporting on the program’s use of AI, federal database access, and grant cancellations.
Readers can follow the full investigation and document archive here:
DOGE Social Security Scandal – Evidence File and Investigation Hub
This article is part of the ongoing DOGE Social Security Data Scandal investigation, documented in the Dystopeka evidence dossier.
DOGE Investigation Status
Status: ACTIVE
Program Involved: Department of Government Efficiency (DOGE)
Programs Affected: Federal DEI, LGBTQ, and humanities grants
Key Evidence: 2026 DOGE deposition testimony from staffers Justin Fox and Nathan Cavanaugh
Last Updated: March 2026
Related Coverage in The Dystopeka Files:
Investigation Summary: DOGE Social Security DEI Scandal
Issue:
Eyewitness deposition testimony reveals that DOGE staff employed AI without expertise, leading to indiscriminate and biased DEI grant terminations.
Key Testimony:
Justin Fox and Nathan Cavanaugh describe reliance on AI, including misclassification of “DEI” projects.
Consequences:
1,400 federal grants terminated; real harm to marginalized communities.
Allegations of Bias:
Homophobic and racist language and decisions cited in depositions.
Status:
Ongoing scrutiny, backlash from advocacy groups and scholars.
DOGE Deposition Timeline: Social Security DEI Scandal
Launch of DOGE Program
The Department of Government Efficiency (DOGE) was launched, with the mission to streamline government operations and reduce spending, including federal grant reviews.
AI Implementation in Grant Reviews
AI tools, including ChatGPT, were introduced by DOGE to analyze federal grants, with the goal of cutting funding to programs deemed unnecessary. The first “DEI-related” programs flagged were diversity initiatives and LGBTQ-focused education.
First Whistleblower Reports
Whistleblowers within DOGE reported that grant terminations were being made based on keywords such as “equity,” “diversity,” and “LGBTQ,” without proper context or understanding of the programs.
Deposition of Justin Fox and Nathan Cavanaugh
In their depositions, DOGE employees Justin Fox and Nathan Cavanaugh admitted that none of the cuts they made actually resulted in budget savings or a reduction in the deficit, revealing that the AI-driven decisions were ineffective.
Public Outcry and Legal Actions
Public backlash against the cuts intensified, especially from marginalized groups and educational institutions. A class-action lawsuit was filed, accusing DOGE of discriminatory and unlawful decision-making.
Congressional Oversight
Congressional hearings were initiated to investigate the allegations of AI misuse and the potential violation of civil rights in the termination of grants. The hearings have raised questions about the role of AI in federal governance.
DOGE Deposition: “AI Does Not Know What DEI Means”
Let’s set the stage: Justin Fox and Nathan Cavanaugh, two young DOGE employees with zero background in government operations, were handed the Herculean task of reviewing thousands of NEH (National Endowment for the Humanities) grants. With little more than ChatGPT and a vague memo about diversity, equity, and inclusion (DEI), they were tasked with cutting projects that seemed “suspect.”
“I ran these project descriptions through ChatGPT, and it basically told me that if the term ‘LGBTQ’ was in the project, it’s probably DEI,” Fox explained, clearly more confused than confident.
The results were, predictably, disastrous.
Watch the DOGE Deposition Videos
Select a video to watch from the dropdown menu below:
Video 1: Doge Staffer Justin Fox Struggles to Define DEI and How They Used AI to Cut LGBTQ Grants
DOGE Deposition Reveals Harm to DEI and LGBTQ Programs
Despite Justin Fox’s admission that none of the cuts made by DOGE’s AI system actually reduced the budget or the deficit, the decision to terminate 1,400 grants devastated crucial programs. Many of these grants were aimed at marginalized communities, including DEI, LGBTQ, and other equity-focused initiatives. The cuts were driven by **homophobic ideologues** who used broad AI filters based on keywords like “equity,” “diversity,” and “LGBTQ,” without understanding the true purpose of the projects.
The real consequence of these cuts was the loss of funding for educational programs on Black history, LGBTQ+ studies, and Native American cultural preservation. These were not budget-saving actions — they were ideologically motivated attacks on initiatives meant to promote inclusion and understanding.
DOGE Deposition Testimony Shows AI Was Used Without Expertise
Fox and Cavanaugh relied on a tool that was, according to them, supposed to be a “helpful intermediary” between their inability to define DEI and the seemingly endless pile of grants they needed to terminate. This led to a list of the most absurd grant cuts imaginable, including Holocaust survivor documentaries and Black history education programs.
“If ChatGPT says it’s DEI,” Cavanaugh said, “then it’s DEI. And we didn’t question it.”
Here’s the kicker: Fox could barely define DEI beyond referencing former President Trump’s executive order banning certain diversity initiatives.
“It’s just… you know, DEI stuff. Equity. Inclusion. Diversity. I don’t know, man.”
Well, at least he had the buzzwords down.
Misfires from Keyword-Based Reviews: Not Just Humanities Grants
One frequently shared comment under the deposition videos claimed that DOGE even cancelled a U.S. Forest Service job responsible for ensuring healthy ecological diversity in forests — allegedly because the word “diversity” appeared in the job’s title. No credible reporting or court documents confirm that specific cancellation.
However, multiple investigations have shown that DOGE’s review process relied too heavily on keyword triggers (like “equity” or “diversity”), leading to unrelated programs being flagged for termination. Critics likened it to a “find and replace” sweep: if the word was there, it was targeted — regardless of actual mission or meaning.
For example, federal programs and manuals have been flagged or removed simply because they used terms like “gender,” “equity,” or other language now associated with DEI, even though their intent was unrelated to DEI initiatives.
This pattern underscores how DOGE’s methodology — driven by AI and superficial text matching — mischaracterized numerous policies and roles, fueling claims of overreach and misunderstanding of the very concepts it sought to eliminate.
The Impact: DOGE Deposition, A Government Horror Story
As a result of these deeply flawed decisions, 1,400 grants were terminated, amounting to tens of millions of dollars. The backlash was swift, with critics calling it nothing short of a federal blunder.
The DEI Backlash: Real-world Consequences
This wasn’t just an issue for the researchers and grant recipients—it had real-world consequences for marginalized communities. Imagine being a researcher working on a Native American history project and learning that the government cut your funding because ChatGPT flagged your grant as “too DEI.”
While Musk and the DOGE team claim that their AI tools were designed to be impartial, it’s hard to see how bias didn’t creep into the process when AI had the final say. Some experts argue that the real problem wasn’t AI per se, but the fact that DOGE didn’t even know what it was trying to accomplish.
“It’s like hiring a robot to decide who gets a scholarship without ever explaining what the scholarship is for,” said Dr. Margaret Bright, a political science professor at Berkeley.
The DEI Bias: Homophobic and Anti-LGBTQ Actions in Grant Cuts
One of the most alarming revelations came when the depositions revealed the homophobic and anti-LGBTQ biases that influenced the decision to terminate certain grants.
Fox: “I was told that if something had LGBTQ-related words or focused on those communities, we needed to pull the funding.”
Cavanaugh: “Yeah, especially if it was in an educational context. We cut grants that had even the mention of LGBTQ history, because it didn’t seem ‘fiscally responsible’ to some people.”
The homophobic bias was evident in the language used during their evaluation. Several grants focused on LGBTQ studies, transgender rights, and gay history were eliminated because they were seen as “politically motivated” or “socially divisive” under the AI-guided framework.
“Any project dealing with LGBTQ issues was labeled ‘divisive’ by the system,” Fox admitted. “And if we weren’t careful, we’d be accused of funding ‘too many of those projects.’”
ChatGPT’s Decision to Delete LGBTQ Projects
LGBTQ-focused education, for example, was among the projects that ChatGPT flagged as “too controversial” or “in the gray zone” and recommended for deletion. Fox, a self-proclaimed tech enthusiast, said that at no point did they even consider the ethical implications of relying on AI to define what constitutes a harmful or divisive project. Cavanaugh chimed in:
“In hindsight, it’s pretty clear we were using AI like a hammer, and everything looked like a nail. I didn’t even realize we were cutting things that should have been fundamental programs for the community.”
No Deficit Reduction: Just AI-Driven Waste
But here’s the kicker: After cutting 1,400 grants, including many related to marginalized communities and LGBTQ rights, Fox and Cavanaugh admitted that none of it reduced the deficit. Not one cent was saved, and the only result was a widespread backlash and public outcry.
Fox: “We didn’t save any money. Nothing changed. But we were told to look busy, and AI seemed like the easiest way to make it look like we were doing something.”
Cavanaugh: “It felt like it was a distraction, honestly. There was no clear goal, and no one ever explained the connection to reducing the deficit.”
The government’s supposed attempt to streamline operations ended up creating greater waste, without any measurable benefit to taxpayers. The cuts were not made for the sake of saving money, but to push a political agenda through AI—and the public suffered for it.
Legal Fallout: The Dogged Lawsuit
Of course, no deposition would be complete without a lawsuit. A class-action lawsuit has now been filed by the Modern Language Association and American Historical Association, claiming that the DOGE team’s actions were not only reckless but also discriminatory. The suit argues that AI-driven bias in grant cuts violated civil rights protections and disproportionately impacted scholars from underrepresented groups.
“They used a machine-learning model to decide whose work mattered,” the lawsuit states. “This is algorithmic discrimination at its finest.”
It’s clear that DOGE didn’t fully understand the implications of letting AI play such a central role in the decision-making process. The testimony raised serious questions about the ethics of using AI to make cultural and educational decisions without human oversight.
Evidence Index
Evidence Index
01
Background
DOGE initiative launch and structure
Early reporting and records describing the creation of the Department of Government Efficiency and concerns about access to federal systems.
02
Whistleblower
Whistleblower allegations regarding SSA access
Reports alleging that DOGE personnel accessed or handled sensitive Social Security databases.
03
Oversight
Congressional oversight inquiries
Members of Congress requested briefings and oversight following whistleblower allegations.
04
Watchdog
SSA watchdog investigation
The Social Security Administration Office of Inspector General opened a formal investigation into possible misuse of SSA databases.
05
Dystopeka Files
Dystopeka investigative reporting series
Ongoing investigative coverage documenting the DOGE program, whistleblower allegations, and federal oversight.
What Went Wrong? The “AI” Problem is the Human Problem
If there’s one thing we learned from the DOGE fiasco, it’s this: AI is only as good as the people who wield it. When Musk’s Department of Government Efficiency handed critical decisions to a language model that couldn’t even spell “equity” correctly half the time, it was bound to be a mess.
Here’s where the real tragedy lies: It wasn’t just about grant cuts or bureaucratic ineptitude—it was about the lack of expertise in something as important as federal funding for humanities and marginalized groups.
The Takeaway: Don’t Let AI Do Your Thinking
In the end, the DOGE deposition serves as a warning to all who are tempted to use AI as a shortcut to decision-making. The key to efficiency lies not in blindly trusting the machine, but in making sure humans with the necessary expertise are still in the driver’s seat.
And, until that happens, the world of government efficiency will remain just another laughingstock in the age of AI.
Archive of investigative reporting examining government power structures and systemic corruption.
DOGE Files,The Dystopeka Files Investigation Series Archive
Additional investigative reporting and related articles.
View the Dystopeka Investigations IndexThis article is connected to newer related coverage and supporting context below.
🔁 Keep Reading
Related Dystopeka reporting and context:
