Rewiring through AI's Social Security Administrators has begun, training videos are not expected
Elon Musk's doe has tried to dig out the federal bureaucracy in the United States in recent months. After the efforts to fire workers and attempt to reduce prestigious institutions, billionaires-backed efforts have sought to create a new governance model that automates priorities. On Friday, Wired reported that in an attempt to modernize the agency, new Chatgpt-style robots were integrated into the workflow of agent staff.
In an internal email viewed in the magazine, the agency supports partners who should “assist employees with daily tasks and increase productivity.” However, the chatbot doesn't seem to work very well. “Honestly, no one is really talking about it,” a source working at the agency told Wired.
The release of the app was reportedly accompanied by an interesting training video (see here) involving an animation, a four-fingered woman. The video was supposed to explain to staff how to use the app’s video, which ignores a very critical message: they should avoid uploading sensitive personal information to the program. The agency forced the agency to send an apology email to staff that highlighted the missing context: “We apologize for the oversight in the training video,” it wrote.
“I'm not sure most of my colleagues have watched the training video,” SSA sources told WIRED. “I played with the chatbot and some of the answers I received from it were very vague and/or inaccurate.” They added, “You'll hear my colleagues teasing the graphics. I know no one is here.” [using it]. It's so clumsy and terrible. ”
If Musk's plan is to automate SSA, there is a lot of evidence (besides the horrible new app) that it is a bad idea. Indeed, similar attempts to automate social services in Brazil show why over-reliance on algorithms to run welfare programs may lead to worse results for everyone.
The rest of the world reported that the attempt by the Brazilian government to reduce bureaucracy by replacing officials with algorithms does not always yield the best results. Brazil has an app developed by state-owned company DataPrev, which is designed to handle social security claims. The application, launched in 2018, uses computer vision and natural language processing to analyze documents submitted by claimants to the government. Unfortunately, the app is used to rejecting legitimate claims based on small mistakes. These automated decisions could initiate a long legal battle, which would take months to resolve.
The export records the experience of a former sugar cane worker, 55, who applied for her retirement benefits through the app but was mistaken for a person by the automation system and denied her benefits. “I have all the documents that prove my health, prove everything, [the benefit] Still rejected. It was a humiliation,” Derito said.
Media noted that rural migrant workers in the country are struggling with the digital face of social services. “The people here can’t [even] Work with Gmail, Facebook, Instagram,” Francisco Santana, president of the Rural Workers Alliance at Barra Do Corda, told Other parts of the world. “The process is [getting] More and more automation and societies are not ready for people living in rural areas, especially in suburban areas. ”
The struggles in automation in other countries are a potential warning to the United States, which is currently “modernizing” technical officers and soldiers slaves deployed by the Trump administration. Musk's Governor attempted to conduct business in the entire federal bureaucracy, including the SSA. Doge replaced full-time employees and tried to develop a “AI-First” strategy that would cut the federal workforce in half and replace human workers with software. A person with a lot of knowledge about Doge's activities recently learned about The Washington Post. However, so far, changes in Doge have not improved efficiency, stimulating dysfunction and confusion in many institutions, including social security.
Indeed, the recent incident in SSA, in which it was reported that, in that incident, the incident reportedly marked “numbers” of life welfare recipients as “death” and cut off their welfare, is just an example. “[DOGE staffers] Entering the system, they killed people,” Rennie Glasgow, a long-time claimed technical analyst for the agency, told Daily Beast. “About 4 million people tagged them as dead. But they weren't sure if these people should be marked as dead, so they sent us an email saying, “If these people come into the office as they are, you can recover them.”
Glasgow added: “We have to go through this long process to resurrect them and keep them alive, which can take three to four days.”
Doge also recently announced that it will try to rewrite the entire code base of SSA in a few months. Sources suggest that at the thrilling pace of their plan to work, Doge almost certainly has to rely on AI to do this. While writing code using AI has become a more common practice, it also requires a lot of supervision due to the extent to which software makes mistakes. Consider how many errors there are Human The operator made a reason during the blitz against the government, believing that automatic rewriting of the SSA code base could be a disaster.
Of course, some people speculate that making mistakes is the key. These critics argue that Doge is not trying to “modernize” or improve SSA, but is actually trying to destroy and destroy the agency so that it can be privatized. If so, everything will be planned.
