JBizNews Desk | May 10, 2026
A federal judge delivered one of the strongest legal rebukes yet against the Department of Government Efficiency Thursday evening, issuing a sweeping 143-page ruling that blocked DOGE’s mass cancellation of humanities grants and sharply criticized the agency’s use of ChatGPT to help determine which federally funded programs should be eliminated.
The ruling by U.S. District Judge Colleen McMahon found the grant terminations unconstitutional and concluded DOGE officials lacked legal authority to direct the cuts in the first place.
The decision is being viewed as a major legal setback not only for DOGE itself, but also for the broader use of artificial intelligence in government decision-making.
What DOGE Actually Did
The lawsuit was brought by the American Council of Learned Societies, which challenged DOGE’s termination of more than $100 million in grants distributed through the National Endowment for the Humanities.
Court filings revealed that DOGE staffers Justin Fox and Nate Cavanaugh used ChatGPT to help identify grants they believed related to DEI — diversity, equity, and inclusion — initiatives.
Those grants were then flagged for cancellation.
According to the ruling and supporting documents, the process led to significant errors.
In one widely cited example, a museum lost a $349,000 federal grant intended to replace its HVAC heating and cooling system after ChatGPT reportedly flagged the proposal as DEI-related.
The project had nothing to do with diversity programming.
The AI system appears to have associated certain language in the application with DEI terminology and incorrectly categorized it.
That mistake became one of the clearest examples cited by critics warning about the risks of using generative AI systems in high-stakes government decisions.
Judge McMahon’s Opinion Was Blunt
Judge McMahon’s ruling did not merely reverse the cuts — it openly questioned the legality and competence of the process itself.
The judge concluded:
- DOGE lacked constitutional authority to terminate congressionally appropriated funding
- The grant cancellations violated separation-of-powers principles
- Congress, not executive agencies, controls federal spending authority
- AI-assisted decision-making without proper oversight created unacceptable legal and operational risks
The opinion represents one of the first major federal rulings directly examining how generative AI tools were used inside government operations.
And the court appeared deeply troubled by what it found.
The Depositions Made the Situation Worse
Public scrutiny intensified after deposition videos from DOGE officials circulated online during the litigation.
During questioning, DOGE staffer Nate Cavanaugh was asked whether he regretted that organizations and workers lost funding and income because of the cuts.
His response:
“No.”
Cavanaugh argued that reducing the federal deficit was more important.
An attorney then asked:
“Did you reduce the federal deficit?”
The exchange quickly went viral and became symbolic of broader criticism surrounding DOGE’s aggressive cost-cutting tactics.
Judge McMahon herself reportedly expressed skepticism during hearings when government attorneys later sought to remove the videos from circulation.
“Are they not proud of what they did?” the judge reportedly asked from the bench.
Why This Case Matters Beyond Humanities Grants
Legal experts say the ruling carries implications far beyond the National Endowment for the Humanities.
At the center of the decision is a constitutional issue:
Who has the legal authority to cancel federal spending already approved by Congress?
Judge McMahon concluded that DOGE’s actions effectively attempted to override congressional appropriations through executive action — something courts have historically treated with extreme caution.
That reasoning could potentially affect:
- Other DOGE-directed funding cuts
- Future executive spending disputes
- AI-assisted federal administrative actions
- Broader questions surrounding executive authority
The ruling also places a major spotlight on the growing use of consumer AI systems inside government operations.
The AI Problem at the Center of the Case
Perhaps the most consequential aspect of the ruling involves the use of ChatGPT itself.
Administrative law scholars and technology experts have repeatedly warned that generative AI systems:
- Can hallucinate facts
- Misclassify information
- Produce inaccurate summaries
- Generate false confidence around uncertain conclusions
Those risks become far more serious when the systems are used to make decisions involving:
- Federal funding
- employment
- legal rights
- public services
- regulatory enforcement
The HVAC grant example became especially damaging because it illustrated how AI errors can directly affect real institutions, workers, and communities.
Bridget Dooling, an administrative law expert and former Bush administration official, described the DOGE approach as “the most risky version of AI that could be applied to regulations.”
The Broader AI Governance Debate Is Now Here
The ruling lands at a moment when governments and corporations across the world are rapidly integrating AI tools into operations.
Many organizations have embraced generative AI for:
- document review
- customer service
- compliance
- budgeting
- hiring
- policy analysis
But the DOGE case may become one of the clearest warnings yet about what happens when AI systems are used without sufficient:
- human oversight
- legal safeguards
- transparency
- accountability
For businesses, the implications are significant.
If courts begin scrutinizing AI-assisted decision-making more aggressively, companies relying heavily on automated systems for consequential actions may face:
- litigation risk
- compliance challenges
- regulatory scrutiny
- reputational damage
What Happens Next
The Trump administration is expected to appeal the ruling.
The case could move to the Second Circuit Court of Appeals and potentially reach the Supreme Court, which has already weighed in this year on broader disputes involving executive authority and federal powers.
For now, the ruling temporarily blocks the grant terminations and opens the possibility that some organizations may eventually recover funding.
But many affected institutions have already:
- laid off staff
- canceled programs
- delayed projects
- reduced operations
And regardless of how the appeals process unfolds, the decision may already have established something larger:
A federal judge has now formally warned that using AI systems to make sweeping government decisions without proper authority or oversight is not simply risky.
It may also be unconstitutional.
© JBizNews.com. All rights reserved. This article is original reporting by JBizNews Desk. Unauthorized reproduction or redistribution is strictly prohibited.
