Top News

A new federal lawsuit has raised concerns about the use of artificial intelligence in government decision-making after academic organizations alleged that a federal initiative relied on the AI chatbot ChatGPT to help identify and cancel grants linked to diversity, equity, and inclusion (DEI) programs.

The case focuses on a government review process introduced during the administration of Donald Trump. According to the complaint, officials used AI tools to analyze grant proposals and determine whether they could be categorized as supporting DEI efforts. Grants identified in that way were then considered for termination.

The lawsuit was filed by two prominent academic groups, the American Council of Learned Societies and the American Historical Association. The organizations argue that the review process resulted in the cancellation of several humanities grants and may have bypassed the traditional expert evaluation system used by federal agencies.

At the center of the dispute is a government office called the Department of Government Efficiency, commonly referred to as DOGE, which was created to examine federal spending across agencies.

Federal Spending Review and the Creation of DOGE

The Department of Government Efficiency was established by executive order in January 2025 as part of a broader effort to reassess federal spending programs. The office was tasked with reviewing budgets and projects funded by various government agencies, including grants awarded through the National Endowment for the Humanities (NEH).

The NEH is responsible for distributing federal funding to support research, preservation projects, museums, and educational initiatives to history, culture, and the humanities. Grant proposals submitted to the agency typically undergo a peer-review process involving scholars and experts who evaluate the projects based on academic merit and public benefit.

According to the lawsuit, the Department of Government Efficiency introduced a new review process aimed at identifying grants tied to diversity, equity, and inclusion initiatives. As part of that effort, staff members reportedly turned to the AI system ChatGPT to analyze descriptions of funded projects.

AI Analysis of Grant Descriptions

Documents cited in the legal filing suggest that employees working for DOGE used the chatbot to review summaries of grant proposals. The staff reportedly asked the AI system to assess whether particular projects were to DEI themes.

A deposition included in the case from a DOGE staff member named Justin Fox describes how the process worked. According to the testimony, employees entered descriptions of grant proposals into the AI tool and asked whether the projects were connected to diversity, equity, or inclusion efforts.

The chatbot’s responses and explanations were then recorded in a spreadsheet maintained by staff. The spreadsheet became a reference point used by the department when deciding which grants should be canceled during the broader spending review.

The lawsuit claims that this spreadsheet eventually replaced an earlier list prepared by experts at the National Endowment for the Humanities. That earlier list had been compiled through the agency’s standard evaluation process.

Critics of the AI-assisted approach argue that this shift represented a significant departure from the traditional peer-review system used for humanities funding.

Museum Grant Became a Key Example

One of the most prominent examples mentioned in the lawsuit involves a grant awarded to the High Point Museum.

The museum had been awarded approximately $349,000 from the National Endowment for the Humanities to upgrade its heating, ventilation, and air-conditioning system. The project was designed to replace aging climate-control equipment that helps preserve artifacts, documents, and other historical items housed at the museum.

Museums rely heavily on stable temperature and humidity levels to prevent deterioration of sensitive materials. According to the museum’s proposal, modernizing the HVAC system would help ensure long-term preservation of the institution’s collection and allow the museum to maintain appropriate conditions for displaying artifacts.

However, the lawsuit claims that when the project description was reviewed using ChatGPT, the system classified the grant as to diversity, equity, and inclusion. The reasoning reportedly focused on the idea that improved preservation conditions could allow broader public access to the museum’s collections.

That classification placed the grant among those identified for cancellation during the federal spending review.

Funding Cut After Work Had Begun

The cancellation occurred after the museum had already begun preparing for the project.

According to reports referenced in the lawsuit, the High Point Museum had started initial planning and early work to the HVAC replacement before the grant was terminated.

Edith Brady, the museum’s director, said the institution was able to recover about 70 percent of the grant’s value through a termination clause in the funding agreement. Even so, the loss of the remaining funds forced the museum to reevaluate how it would proceed with the upgrades.

Museum officials have maintained that the project was focused on preservation and infrastructure improvements rather than diversity programming.

Other Grants Also Reviewed

The lawsuit states that the museum project was not the only grant flagged during the AI-assisted review.

Another example mentioned in the legal complaint involved a proposal submitted by North Carolina Central University, a historically Black university in Durham. The project sought funding to develop educational materials using digital archival collections.

According to the spreadsheet cited in the lawsuit, the proposal was also categorized as connected to DEI themes after being reviewed with the help of ChatGPT.

Projects that received similar classifications were included in the list of grants targeted for termination as part of the government’s review of humanities funding.

Academic Groups Challenge the Process

The organizations that filed the lawsuit argue that the cancellations were unlawful and violated constitutional protections.

The American Council of Learned Societies and the American Historical Association contend that the review process replaced the established system of scholarly evaluation with a faster method that relied heavily on AI-generated interpretations.

According to their complaint, this approach undermined the integrity of the grant-making system administered by the National Endowment for the Humanities.

The lawsuit also claims that terminating grants based on perceived associations with DEI initiatives may have infringed on First Amendment protections by targeting projects for ideological reasons.