Judge Rules DOGE's Use of ChatGPT to Cancel Grants Unconstitutional

In a landmark decision that has sent shockwaves through the federal bureaucracy, a U.S. district judge has ruled that the Department of Government Efficiency's (DOGE) use of ChatGPT to cancel over $100 million in grants was unconstitutional. The 143-page ruling did not mince words, describing the agency's process as both "dumb and illegal." The case centered on DOGE's controversial practice of feeding grant applications into OpenAI's ChatGPT and using the AI's recommendations to determine which programs to defund. According to court documents, the agency bypassed standard review protocols, including peer evaluation and statutory requirements for public notice, in favor of rapid AI-driven cuts. The judge found that this delegation of decision-making authority to an algorithm violated the Administrative Procedure Act and constitutional separation of powers. Legal experts say the ruling establishes a critical precedent for judicial scrutiny of AI in government. "This is a wake-up call for any agency thinking they can outsource their legal obligations to a chatbot," said constitutional law professor Sarah Chen. "The court made clear that AI tools can assist, but cannot replace, human judgment in matters of public funding." The decision also raised alarms about data privacy, as grant applications often contained sensitive personal and financial information. The judge noted that DOGE failed to conduct any privacy impact assessment before feeding data into a third-party AI system. DOGE has announced it will appeal the ruling, arguing that AI tools are necessary to modernize government efficiency. However, the decision has already prompted several federal agencies to pause their own AI pilot programs pending review. For now, the canceled grants must be reinstated, and Congress is considering legislation to explicitly ban AI from making final decisions on federal funding. This case marks a pivotal moment in the ongoing debate over the appropriate role of artificial intelligence in governance, with the judiciary drawing a clear line: algorithms can inform, but they cannot decide.