
Ford School assistant professor (by courtesy) Ben Green, wrote an essay for Tech Policy on Elon Musk's new “AI-first strategy” for the federal government, and commented, "unleashing AI to revolutionize the federal government is a recipe for disaster."
Green says, "There are already many instances of AI-driven governance gone wrong—particularly in cases where the goal is to root out fraud and cut budgets." For example, he highlighted the case of Michigan where the "Unemployment Insurance Agency (UIA) adopted an algorithm (the Michigan Integrated Data Automated System, aka MiDAS) to streamline its operations" and automate a third of the work done by staff. However, he writes "Almost every accusation of fraud—93% of supposed fraud cases—was incorrect. Then, even once these errors became clear, it took years of litigation for the wrongly accused to receive the money they were owed. For many people, the fraud charge remained on their criminal record for years, barring them from jobs."
"It is a typical story," writes Green. "These quick-fix proposals are typically laughable to the existing staff in those agencies, who have deep expertise in navigating the complex nuances of government processes. The engineers do not understand the intricate workflows that government agencies follow." Because of this, DOGE's engineers have struggled to interpret and handle government data and forms, leading to false assumptions and mistakes.
"The most unmistakable evidence that Musk and DOGE are not committed to thoughtfully developing technology that improves government is that they are dismantling the teams that have been pioneers in this work," writes Green. For example, "last weekend, DOGE eliminated 18F, a group that helps federal agencies improve their technology." Green says the key to working in government is "to appreciate the wisdom of existing staff," not replacing them.
Green concludes, "Given the frantic pace and hubris of Musk and the DOGE team, I fear that this AI incursion will do incredible damage to federal operations and the people who depend on them."