The House Office of Legislative Counsel is currently dealing with a growing problem. Staffers and outside groups increasingly rely on generative AI tools to draft legislative language, but the results are causing significant friction. Rather than saving time, these AI-generated proposals arrive riddled with errors, incorrect citations, and poor legal terminology.
Legal experts note that legislative drafting requires a high level of precision. Even minor mistakes, such as misidentifying a grant versus a tax credit, can have massive consequences for the validity of new laws. Because these drafts contain so many errors, the Office of Legislative Counsel now spends more time fixing and rewriting AI-generated text than it would have taken to draft the legislation from scratch.
Congressional staffers report that AI use has become a standard, if often quiet, part of the process. The shift is creating a backlog, as the office works to maintain the quality control necessary for the American legal system. Some experts warn that this trend threatens to undermine the core work of Congress, as the reliance on chatbots prevents staff from deeply learning the issues they seek to regulate.
While some offices are exploring specialized, vetted technology to assist with research, the current reliance on general-purpose chatbots creates risks. As the use of AI in bill drafting grows, the Office of Legislative Counsel faces an uphill battle to keep flawed language out of federal statutes. Lawmakers are now being warned that their ultimate responsibility for the quality of introduced bills cannot be delegated to an automated tool.