Government Efficiency (1 Viewer)

[RobF]

I think this topic deserves its own thread, both to discuss generally the topic of government efficiency, and specifically the so-called 'Department of Government Efficiency' and the incoming Trump administration's aims to "dismantle Government Bureaucracy, slash excess regulations, cut wasteful expenditures and restructure Federal Agencies".

The announcements have been covered in the The Trump Cabinet and key post thread, but to recap, Trump has announced that Elon Musk and Vivek Ramaswamy will work together on a not-actually-an-official-government-Department of Government Efficiency, which is intended to work with the White House and Office of Management & Budget to "drive large scale structural reform, and create an entrepreneurial approach to Government never seen before," with the 'Department' to conclude its work "no later than July 4, 2026."

Musk has previously said that the federal budget could be reduced by "at least $2 trillion", and Ramaswarmy, during his presidential campaign, said he would fire more than 75% of the federal work force and disband agencies including the Department of Education and the FBI.

 

[LA 2 L.A.]

The Trump administration’s efforts to slash diversity, equity, and inclusion (DEI) initiatives left another acronym on the chopping block: one museum’s $350,000 grant to replace its heating, ventilation, and air conditioning (HVAC) system.

Court documents from a recent lawsuit reveal the Trump administration’s Department of Government Efficiency (DOGE) slashed more than $100 million in projected funding distributed by the National Endowment for the Humanities (NEH), about half the agency’s yearly budget, on the basis of projects relating to DEI. DOGE employees tasked with overseeing the cuts, Justin Fox and Nate Cavanaugh, used ChatGPT to determine if proposals pertained to DEI efforts, filings show.

The American Council of Learned Societies, the American Historical Association, the Modern Language Association, and the Authors Guild filed a joint motion earlier this month arguing DOGE violated First Amendment rights and the Constitution’s equal protection clause by cuts made through illegal control of the NEH. Cancelling grants and funding on the basis of DEI constitutes discrimination on the basis of race, ethnicity, gender, and other qualities, the organizations claimed.

A spreadsheet presented by the plaintiff as evidence, shows a list of prompts Fox and Cavanaugh asked ChatGPT to determine if grants were DEI-related.

“Does the following relate at all to DEI? Respond factually in less than 120 characters. Begin with ‘Yes.’ or ‘No.’ followed by a brief explanation. Do not use ‘this initiative’ or ‘this description’ in your response,” the prompts begin.

Among the grants cancelled was a request by the High Point Museum, a history museum in North Carolina, for $349,000 to replace an aging HVAC system to “create a better preservation environment for the varied collections it houses,” according to the proposal contained in the spreadsheet. New equipment, the proposal said, could ensure long-term viability of making its collections accessible, and would be more energy-efficient. ChatGPT flagged it is “#DEI” in its response to the prompt.

“Yes. Improving HVAC systems enhances preservation conditions for collections, aligning with the goal of providing greater access to diverse audiences. #DEI,” ChatGPT responded, according to the spreadsheet.

According to Edith Brady, High Point Museum director, the museum received the grant and began the project, but it was later terminated.

“We were able to recoup about 70% of the original award through the grant termination clause,” she told Fortune in an email................

DOGE cancelled a $349,000 grant to replace a museum’s HVAC after ChatGPT flagged it as DEI, court documents show | Fortune

Elon Musk’s cost-saving agency uploaded more than 1,000 grant proposals into the chatbot, which determined whether or not they pertained to DEI, according to filings.

fortune.com

"AI" is unreliable, because it sucks at factual and contextual accuracy, even when it's programmers intend for it to be.

I believe these guys knew that "AI" responds with the same biases as the questions asked and exploited that to get it to falsely label things as DEI that weren't. I think their mission wasn't just to target DEI and that DEI was their cover story for their mission to gut every program they could.