Elon Musk’s vision to use artificial intelligence to run the US government has drawn serious criticism from experts, with many arguing it could lead to dangerous consequences. Musk’s Department of Government Efficiency (DOGE) has reportedly fired thousands of government workers and now expects the remaining staff to send weekly emails summarizing their accomplishments. With the flood of emails, Musk plans to rely on AI to sift through the responses and assist in decisions about employee retention. His plan also includes replacing many government employees with AI systems.
The details of the AI systems Musk intends to use remain unclear, sparking concerns among US lawmakers, particularly Democrats, who have called for greater transparency. Experts warn that AI could cause severe problems if not rigorously tested and validated before deployment. Cary Coglianese, a professor at the University of Pennsylvania, expressed skepticism about using AI to decide job terminations, pointing out the risk of mistakes, bias, and other issues. Shobita Parthasarathy, a professor of public policy, emphasized the unknowns surrounding the AI’s training and its ability to make trustworthy decisions.
Despite the warnings, Musk’s push to integrate AI into government operations continues, with significant involvement in areas such as social media monitoring. The US Department of State has already planned to use AI to scan foreign nationals’ social media for potential ties to groups like Hamas, though the specifics of these systems remain opaque.
AI experts caution that improper use of AI in government could lead to significant undetected harms. Coglianese highlighted previous global failures in AI deployments that led to harmful outcomes, such as incorrect welfare benefit denials in the UK and Netherlands. Similarly, Michigan’s AI-driven unemployment fraud detection system wrongly accused thousands of fraud, leading to penalties and even bankruptcies before the state acknowledged the flaw.
AI’s failure to grasp nuances could be especially damaging when used in governmental roles that require specialized knowledge, such as in the Department of Justice or Agriculture. The idea that AI could fully replace workers in such roles is seen as impractical by experts like Coglianese, who pointed out that tasks in government jobs often demand expertise that AI lacks.
The Trump administration, with Musk’s influence, is pushing ahead with AI plans despite these concerns. However, there are efforts to ensure AI is used responsibly. Under the Biden administration, an executive order was issued in 2023 to ensure AI in government is used properly, though it was rescinded by Trump. Schellmann, a professor at NYU, noted this rollback as a setback for responsible AI deployment.
Despite these challenges, experts agree that AI, if developed and used correctly, could offer significant benefits, such as automating repetitive tasks to free workers for more important roles. However, careful planning, transparency, and validation are necessary to prevent AI from being misused and causing harm.
