The US Capitol is currently dealing with a new kind of problem as more lawmakers turn to chatbots to write their proposed laws. According to a recent report, the US House Office of Legislative Counsel is getting flooded with AI-generated bills that are packed with careless errors. Instead of speeding things up, this new habit is actually forcing lawyers to spend extra hours fixing bad drafts.
Lawyers spend extra hours fixing basic mistakes in new bills
More and more congressional offices are using general chatbots like ChatGPT and Claude to write their legislative proposals. While this sounds like a great shortcut, the people who actually have to review these documents are not happy. Lawyers at the Office of Legislative Counsel now say they are spending way more time cleaning up machine-made drafts than they would if they just wrote the bills from scratch.
The problem comes down to the tiny details that matter in lawmaking. These artificial intelligence tools often miss crucial nuances. For example, a chatbot might not understand the difference between a tax credit, a tax deduction, or a simple grant. It might also use the word “state” in a way that accidentally leaves out Washington DC and tribal nations. On top of that, lawyers are catching fictional or incorrect references to older laws buried in the text.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
Staffers miss important details when relying on computer chatbots
This growing reliance on tech is creating another issue behind the scenes. When congressional staffers use an AI program to do the heavy lifting, they do not learn the issues as deeply as they used to. Because a computer spits out the text for them, the authors of these bills often struggle to clearly explain the actual goals of the laws they are trying to pass.
To fight back against this messy workload, the legal office is ironically building its own specialized software. They put together a custom tool that helps visualize how a new proposal will actually change current laws. Unlike a normal chatbot, this custom program will simply return an error if it gets confused, preventing wild hallucinations from sneaking into the final legal text.
Even as Capitol Hill tries to figure out how to handle new technology, it is clear that letting robots write the laws still needs a lot of human supervision.
Discussion