How faulty DOGE AI tool impacted VA contracts
The code, using outdated and inexpensive AI models, produced results with glaring mistakes. For instance, it hallucinated the size of contracts, frequently misreading them and inflating their value. It concluded more than a thousand were each worth $34 million, when in fact some were for as little as $35,000.
The DOGE AI tool flagged more than 2,000 contracts for 'munching.' It's unclear how many have been or are on track to be canceled — the Trump administration's decisions on VA contracts have largely been a black box. The VA uses contractors for many reasons, including to support hospitals, research and other services aimed at caring for ailing veterans.
Get Starting Point
A guide through the most important stories of the morning, delivered Monday through Friday.
Enter Email
Sign Up
VA officials have said they've killed nearly 600 contracts overall. Congressional Democrats have been pressing VA leaders for specific details of what's been canceled without success.
Advertisement
We identified at least two dozen on the DOGE list that have been canceled so far. Among the canceled contracts was one to maintain a gene sequencing device used to develop better cancer treatments. Another was for blood sample analysis in support of a VA research project. Another was to provide additional tools to measure and improve the care nurses provide.
Advertisement
ProPublica obtained the code and the contracts it flagged from a source and shared them with a half dozen AI and procurement experts. All said the script was flawed. Many criticized the concept of using AI to guide budgetary cuts at the VA, with one calling it 'deeply problematic.'
Cary Coglianese, professor of law and of political science at the University of Pennsylvania who studies the governmental use and regulation of artificial intelligence, said he was troubled by the use of these general-purpose large language models, or LLMs. 'I don't think off-the-shelf LLMs have a great deal of reliability for something as complex and involved as this,' he said.
Sahil Lavingia, the programmer enlisted by DOGE, which was then run by Elon Musk, acknowledged flaws in the code.
'I think that mistakes were made,' said Lavingia, who worked at DOGE for nearly two months. 'I'm sure mistakes were made. Mistakes are always made. I would never recommend someone run my code and do what it says. It's like that 'Office' episode where Steve Carell drives into the lake because Google Maps says drive into the lake. Do not drive into the lake.'
Though Lavingia has talked about his time at DOGE previously, this is the first time his work has been examined in detail and the first time he's publicly explained his process, down to specific lines of code.
Lavingia has nearly 15 years of experience as a software engineer and entrepreneur but no formal training in AI. He briefly worked at Pinterest before starting Gumroad, a small e-commerce company that nearly collapsed in 2015. 'I laid off 75% of my company — including many of my best friends. It really sucked,' he said. Lavingia kept the company afloat by 'replacing every manual process with an automated one,' according to
Advertisement
Lavingia did not have much time to immerse himself in how the VA handles veterans' care between starting on March 17 and writing the tool on the following day. Yet his experience with his own company aligned with the direction of the Trump administration, which has embraced the use of AI across government to streamline operations and save money.
Lavingia said the quick timeline of
Under a time crunch, Lavingia said he finished the first version of his contract-munching tool on his second day on the job — using AI to help write the code for him. He told ProPublica he then spent his first week downloading VA contracts to his laptop and analyzing them.
VA press secretary Pete Kasperowicz lauded DOGE's work on vetting contracts in a statement to ProPublica. 'As far as we know, this sort of review has never been done before, but we are happy to set this commonsense precedent,' he said.
The VA is reviewing all of its 76,000 contracts to ensure each of them benefits veterans and is a good use of taxpayer money, he said. Decisions to cancel or reduce the size of contracts are made after multiple reviews by VA employees, including agency contracting experts and senior staff, he wrote.
Advertisement
Kasperowicz said that the VA will not cancel contracts for work that provides services to veterans or that the agency cannot do itself without a contingency plan in place. He added that contracts that are 'wasteful, duplicative or involve services VA has the ability to perform itself' will typically be terminated.
Trump officials have said they are working toward a
The VA has said it would avoid cutting contracts that directly impact care out of fear that it would cause harm to veterans. ProPublica recently reported that relatively small cuts at the agency have already
The VA has not explained how it plans to simultaneously move services in-house, as Lavingia's code suggested was the plan, while also slashing staff.
Many inside the VA told ProPublica the process for reviewing contracts was so opaque they couldn't even see who made the ultimate decisions to kill specific contracts. Once the 'munching' script had selected a list of contracts, Lavingia said he would pass it off to others who would decide what to cancel and what to keep. No contracts, he said, were terminated 'without human review.'
'I just delivered the [list of contracts] to the VA employees,' he said. 'I basically put munchable at the top and then the others below.'
VA staffers told ProPublica that when DOGE identified contracts to be canceled early this year — before Lavingia was brought on — employees sometimes were given little time to justify retaining the service. One recalled being given just a few hours. The staffers asked not to be named because they feared losing their jobs for talking to reporters.
Advertisement
According to one internal email that predated Lavingia's AI analysis, staff members had to respond in 255 characters or fewer — just shy of the 280 character limit on Musk's X social media platform.
Once he started on DOGE's contract analysis, Lavingia said he was confronted with technological limitations. At least some of the errors produced by his code can be traced to using older versions of OpenAI models available through the VA — models not capable of solving complex tasks, according to the experts consulted by ProPublica.
Moreover, the tool's underlying instructions were deeply flawed. Records show Lavingia programmed the AI system to make intricate judgments based on the first few pages of each contract — about the first 2,500 words — which contain only sparse summary information.
'AI is absolutely the wrong tool for this,' said Waldo Jaquith, a former Obama appointee who oversaw IT contracting at the Treasury Department. 'AI gives convincing looking answers that are frequently wrong. There needs to be humans whose job it is to do this work.'
Lavingia's prompts did not include context about how the VA operates, what contracts are essential or which ones are required by federal law. This led AI to determine a core piece of the agency's own contract procurement system was 'munchable.'
At the core of Lavingia's prompt is the direction to spare contracts involved in 'direct patient care.'
Such an approach, experts said, doesn't grapple with the reality that the work done by doctors and nurses to care for veterans in hospitals is only possible with significant support around them.
Advertisement
Lavingia's system also used AI to extract details like the contract number and 'total contract value.' This led to avoidable errors, where AI returned the wrong dollar value when multiple were found in a contract. Experts said the correct information was readily available from public databases.
Lavingia acknowledged that errors resulted from this approach but said those errors were later corrected by VA staff.
In late March, Lavingia published a version of the 'munchable' script
According
That openness may have eventually led to Lavingia's dismissal. Lavingia confirmed he was
VA officials have declined to say whether they will continue to use the 'munchable' tool moving forward. But the administration may deploy AI to help the agency replace employees. Documents previously obtained by ProPublica show DOGE officials proposed in March consolidating the benefits claims department by relying more on AI.
And the government's contractors are paying attention. After Lavingia posted his code, he said he heard from people trying to understand how to keep the money flowing.
'I got a couple DMs from VA contractors who had questions when they saw this code,' he said. 'They were trying to make sure that their contracts don't get cut. Or learn why they got cut.
'At the end of the day, humans are the ones terminating the contracts, but it is helpful for them to see how DOGE or Trump or the agency heads are thinking about what contracts they are going to munch. Transparency is a good thing.'
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Wall Street Journal
30 minutes ago
- Wall Street Journal
Treasury Yields Rise on Stable Employment Ahead of CPI
1600 ET – U.S. job creation slows less than expected, reducing odds of a dovish Fed. Bond markets react with a selloff that boosts yields. May's job creation slows less than forecast and unemployment remains at 4.2%. CME data show diminishing odds of a rate cut before September. Two or more cuts this year still represent the highest odds, but bets on only one or no cut rise. Wells Fargo foresees May's 12-month core CPI, due Wednesday, accelerating to 3.3% from April's 2.8%. The 10-year gains 0.089 percentage point this week, including 0.155 p.p. today, to 4.507%. The two-year rises 0.125 p.p. in the week and 0.115 p.p. today, to 4.039%. ( @ptrevisani) 0846 ET – U.S. job creation didn't slow as much as expected in May, spurring a bonds selloff that takes Treasury yields higher. May payrolls slowed to 139,000 from a downwardly revised 147,000. Economists surveyed by WSJ forecast 125,000. Unemployment was unchanged at 4.2%, as expected. The data likely supports expectations of a Fed hold. Yields were already rising ahead of payrolls, as markets watched the Trump-Musk break up. They rose faster after the data, particularly in longer maturities. The 10-year trades at 4.452%% and the two-year at 3.985%. ( @ptrevisani)
Fox News
33 minutes ago
- Fox News
Lara Trump: China wants to replace us
All times eastern FOX News Saturday Night with Jimmy Failla FOX News Radio Live Channel Coverage WATCH LIVE: President Trump attends UFC event in Newark, New Jersey
USA Today
33 minutes ago
- USA Today
4 Social Security changes Washington could make to prevent benefit cuts
4 Social Security changes Washington could make to prevent benefit cuts Show Caption Hide Caption Biden criticizes Trump administration's handling of Social Security Social Security overhaul sparks criticism from Biden over service disruptions, layoffs and automation as Trump defends changes as efficiency. Straight Arrow News Social Security is an important source of income for millions of Americans, but the program has a serious financial problem. Costs have increased faster than revenues in recent years because the aging population is growing more quickly than the working population. As a result, the trust fund, the financial account that pays benefits, is on track to be depleted within a decade. Specifically, the Congressional Budget Office estimates the trust fund will be exhausted in 2034. That would eliminate one source of revenue (i.e., interest earned on trust fund reserves), and the remaining tax revenues would only cover 77% of scheduled payments. That means a 23% benefit cut would be necessary in 2035. Fortunately, the lawmakers in Washington have several years to find a better solution. Here are four Social Security changes that could prevent deep, across-the-board benefit cuts. 1. Apply the Social Security payroll tax to income above $400,000 Social Security is primarily funded by a dedicated payroll tax, which takes 6.2% of wages from workers and employers. But some income is exempt from the payroll tax. Specifically, the maximum taxable earnings limit is $176,100 in 2025. Income above that threshold is not taxed by Social Security. Importantly, the Social Security program is projected to run a $23 trillion deficit over the next 75 years as it's strained by shifting demographics. But the deficit could be slashed by applying the payroll tax to more income. For instance, including income above $400,000 would eliminate 60% of the 75-year funding shortfall, says the University of Maryland. 2. Gradually increase the Social Security payroll tax rate to 6.5% over six years Under current law, the Social Security payroll tax rate is 6.2% for workers and their employers. But gradually raising that figure would eliminate a portion of the long-term deficit. For example, increasing thetax rate by 0.05% annually over a six-year period would eliminate 15% of the 75-year funding shortfall, according to the University of Maryland. Now that I've discussed two possible changes, let's step back and look at the big picture. There are basically three ways to resolve Social Security's financial problems: (1) increase revenue, (2) reduce costs or (3) some combination of the first two options. The changes discussed so far would increase revenue, but the next two changes would cut benefits. However, they are more subtle cuts than the 23% across-the-board reduction that would follow trust fund depletion. 3. Gradually increase full retirement age to 68 by 2033 Workers are eligible for retirement benefits at age 62, but they are not entitled to their full benefit — also called the primary insurance amount (PIA) — until full retirement age (FRA). Anyone that claims before full retirement age receives a smaller payout, meaning they get less than 100% of their PIA. FRA is currently defined as 67 years old for workers born in 1960 or later, but raising the figure would reduce the long-term deficit. For instance, increasing FRA to 68 years old by 2033, meaning it would apply to workers born in 1965 or later, would eliminate 15% of the 75-year funding shortfall, according to the University of Maryland. 4. Reduce benefits for retired workers with income in the top 20% Social Security benefits are determined as percentages of two bend points. Specifically, income from the 35 highest-paid years of work is adjusted for inflation and converted to a monthly figure called the average indexed monthly earnings (AIME) amount. The AIME is then run through a formula that uses two bend points to determine the PIA for each worker. Modifying the second (highest) bend point would eliminate a portion of the long-term deficit by reducing benefits for high earners. For instance, the University of Maryland estimates that reducing benefits for individuals with income in the top 20% could reduce the 75-year funding deficit by 11%. Here's the big picture: The four changes I've discussed would eliminate 101% of Social Security's $23 trillion funding shortfall, which would prevent across-the-board benefit cuts in 2035. The Motley Fool has a disclosure policy. The Motley Fool is a USA TODAY content partner offering financial news, analysis and commentary designed to help people take control of their financial lives. Its content is produced independently of USA TODAY. The $23,760 Social Security bonus most retirees completely overlook Offer from the Motley Fool: If you're like most Americans, you're a few years (or more) behind on your retirement savings. But a handful of little-known "Social Security secrets"could help ensure a boost in your retirement income. One easy trick could pay you as much as $23,760 more... each year! Once you learn how to maximize your Social Security benefits, we think you could retire confidently with the peace of mind we're all after. JoinStock Advisorto learn more about these strategies. View the "Social Security secrets" »
