Politico reports that the House Office of Legislative Counsel, the office of career attorneys that turns members’ policy ideas into legally precise bill text, is being flooded with AI-generated legislative drafts riddled with errors.
The reporting draws on interviews with eight current and former officials who work (or have worked) with the office. Congressional staffers and outside groups are using Claude and ChatGPT to draft legislative language, and one adviser told Politico the office now spends more time fixing AI-drafted bills than it would take to write them from scratch.
This should surprise absolutely no one. LLMs and reasoning engines (such as ChatGPT, Claude, and Gemini) are simply raw intelligence; without proper specialized knowledge bases and guardrails, they have no way to craft professional-grade legislation.
The article listed several easily predictable errors. Wade Ballou, who led the office from 2016 to 2024, said AI misses details like whether money should flow as a tax credit, a tax deduction, a tax exclusion, or a grant. Drafts can define “state” in ways that accidentally exclude D.C. and tribal nations from federal programs. AI-generated text also routinely miscites the U.S. Code, which lawyer and technologist Ari Hershowitz called “a guaranteed way of introducing thousands of bugs into our legal system.”
Importantly, AI didn’t “miss” anything. Without a proper humans-in-the-loop eval strategy, there is no way for a generic AI product to yield usable results.
So far, no bill voted on by either chamber has been shown to be entirely AI-crafted, but AI’s influence is clearly visible. In June, an amendment summary from Rep. Anna Paulina Luna’s office went viral because a Claude response was pasted into the text. She said her staff used AI to spell-check the summary and that the amendment itself was written without AI. Her assessment: “Not a shocker. Most staff use it.”
Rep. Norma Torres put the institutional position plainly: responsibility for introduced legislation falls on members and their staff, and “that responsibility cannot be outsourced to a chatbot.” With all due respect to Rep. Torres, while this may be true today, there will come a time in the very near future when outsourcing to a “chatbot” (to use her words) may be preferable.
Every company needs a Claw strategy. Do you have one?
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.
About Shelly Palmer
Shelly Palmer is the Professor of Advanced Media in Residence at Syracuse University’s S.I. Newhouse School of Public Communications and CEO of The Palmer Group, a consulting practice that helps Fortune 500 companies with technology, media and marketing. Named LinkedIn’s “Top Voice in Technology,” he covers tech and business for Good Day New York, is a regular commentator on CNN and writes a popular daily business blog. He's a bestselling author, and the creator of the popular, free online course, Generative AI for Execs. Follow @shellypalmer or visit shellypalmer.com.