AI-generated 'slop' creates bottleneck for House legislative lawyers
Lawyers at the Office of Legislative Counsel report that correcting AI-drafted bills often takes longer than writing them from scratch.
The U.S. House Office of Legislative Counsel (OLC) is struggling to keep pace with a surge of AI-generated bill drafts submitted by congressional offices and outside advocacy groups. This influx of automated text is creating a significant bottleneck for the nonpartisan civil servants responsible for turning policy ideas into precise legal language.
OLC lawyers report that reviewing and correcting AI-drafted legislation often takes more time than drafting the bills from scratch. These drafts are frequently riddled with legal errors and a lack of nuance. Common failures include incorrect citations of existing statutes and a fundamental inability to distinguish between different financial mechanisms, such as tax credits, deductions, exclusions, and grants.
The risk of legal 'bugs'
The precision of legislative language is critical to avoid unintended consequences. One specific danger is the way general-purpose AI tools handle definitions. For example, AI-generated definitions of the word "state" may inadvertently exclude the District of Columbia and tribal nations from federal programs by only accounting for the 50 states.
Industry experts warn that this trend risks introducing thousands of "bugs" into the legal system. Daniel Schuman, executive director of the American Governance Institute, told Politico that while AI is useful for many tasks, it is not capable of drafting legislation that should be enacted into law. Rep. Norma Torres echoed this sentiment, stating that the responsibility of drafting law cannot be outsourced to a chatbot.
Erosion of expertise
Beyond the immediate technical errors, there are growing concerns regarding the diminishing subject-matter expertise within the Capitol. Reports indicate that congressional staffers are becoming less familiar with the deep details of their own bills because they rely on AI for the initial drafting process. This shift moves the burden of technical accuracy away from the policy creators and onto the OLC lawyers who must clean up the resulting "slop."
Safeguards and next steps
To combat these inaccuracies, the OLC and the Office of the Clerk utilize the Comparative Print Suite. Launched in 2022, this set of tools is designed to visualize how proposals change current laws. Unlike generative AI, the suite is built for precision; it is designed to return errors when changes cannot be determined rather than hallucinating a plausible but incorrect answer.
While the OLC continues to explore how AI might improve internal efficiencies, the current trend of external AI-drafting remains a point of contention. An anonymous House staffer described the situation as "absolutely terrifying," noting that the broader implications of automated lawmaking may not yet be fully understood by those in power.