What’s worse than DOGE firing a whole bunch of highly skilled government employees? Replacing them with likely bad AI.
One of the kinds of headlines almost everyone in the US is seeing revolves around the badly named organization, DOGE (Department of Government Efficiency) firing thousands of government employees randomly. This, of course, is an action being conducted under the bad thinking that people employed in the government is just government waste.
This has, of course, led to many complaints ranging from ones that deliver a hilarious amount of schadenfreude where MAGA supporters are realizing that the leopards have come to eat their faces. On the other end of the spectrum, we see how the cuts are costing lives and compromising national security. Simply put, the Elon Musk led to random mass firings that is part of the reason why the US is in early stages of total collapse.
Fear not, however. The stable geniuses of the Trump administration leadership have come up with a solution to the loss of government functionality and American society facing total collapse. That solution is to simply replace all the fired employees with an unproven and hastily put together AI. In a roundabout way, it’s almost like an admission that government employees actually do something useful after all. Regardless, if you wonder why the government is telling you to eat 18 gallons of ice cream before applying for unemployment services, at least you’ll know why. For those who are wondering if I’m making all of this up, well, here’s Gizmodo:
Elon Musk and his Department of Government Efficiency are attempting to enact what some experts have called the “largest job cut in American history“—but don’t worry, these geniuses have a solution to pick up the significant amount of slack caused by letting go of tens of thousands of domain experts and civil servants all at once: a chatbot.
According to Wired, DOGE has given about 1,500 employees at the US General Services Administration, the agency that manages federal real estate and oversees most government contracts, access to a proprietary chatbot called GSAi. That’s right, the agency that has already lost hundreds of employees to termination or resignation, including basically everyone working at its extremely efficient tech hub known as 18F, is getting ChatGPT in a suit that matches federal dress code to make up for all that lost labor.
GSAi, which was apparently rushed out the door by DOGE with the intention of deploying it across the entire agency, is supposed to support staff with “general” tasks. In an internal memo obtained by Wired, GSA employees were told that when it comes to what they can use GSAi for, “the options are endless.” It then offered a list of tasks that, frankly, ended very quickly: “You can: draft emails, create talking points, summarize text, write code.”
Employees were also given a pretty major caveat about how they can use GSAi: no nonpublic information or “controlled unclassified information”—information that is sensitive but not classified—can be shared with it. That’s an understandable but pretty limiting disclaimer, especially if an employee wants to use the chatbot to, say, summarize meeting notes or help structure some data. Fittingly, a GSA employee told Wired the chatbot is “about as good as an intern,” and it produces “Generic and guessable answers.”
Indeed, several AI Large Language Modules (LLMs) that generate content have generally been quite terrible with what they do. Whether that is Google’s AI telling people to eat rocks or Apple’s Intelligence AI falsely reporting that Luigi Mangione had shot himself, the ways in which AI is failing miserably is quite a long and extensive list. Either way, when AI that has been in development for years specifically for a very particular is failing badly, it’s hard to envision a hastily put together last minute AI is somehow going to be replacing federal workers who do a huge variety of tasks. After all, government employees do more than just write e-mails and answer questions from the public.
Either way, this is just a recipe for disaster. There is no way I can see that all of this ends well beyond the far right pretending everything is hunky dory. If anything, a case can be made that replacing critical government employees with AI is worse than the critical position being left unfilled. After all, would you rather your car with several issues get left untouched or smashed to bits by someone who clearly doesn’t know what they are doing? Probably the former.
After all, if you are using AI to help you tighten up the grammar in an e-mail, that’s fine. Using AI to fully replace employees, however, is completely different and unlikely to succeed.
