AI-Generated Bill Drafts Flood U.S. House of Representatives, Legislative Counsel Office Overwhelmed
A large number of AI-generated bill drafts are flooding into the U.S. House of Representatives, many of which contain legal errors, inaccurate citations, and overly vague or verbose wording.
The House Office of Legislative Counsel is now spending more time amending these drafts than it would take to draft bills from scratch.
Officials warn that these errors could slow down the legislative process, lead to lawsuits, and make it difficult for administrative agencies and courts to interpret the law.
Congressional offices and external groups are increasingly using generative AI to draft legislative texts, resulting in a surge in workload for the office.
Common issues include incorrect citations of U.S. Code sections, confusion over legal terminology definitions, and unintentional exclusion of specific regions or groups.
Market mechanisms indicate that event-driven legislative efficiency is being reevaluated against compliance risks: AI tools accelerate draft production, but the costs of error correction are rising, increasing the demand for professional legal services and reviews.
Additionally, the office's lawyers and support staff have previously handled over 30,000 draft requests, with AI further increasing the burden.
Source: Public Information
ABAB AI Insight
The House Office of Legislative Counsel has long been responsible for translating policy intentions into precise legal texts. With the rise of generative AI, congressional staff and lobbying groups have begun submitting AI drafts in bulk, shifting the office's role from drafters to primary error correctors.
In terms of capital pathways, AI tools lower the threshold for draft generation, reallocating resources from professional drafting to post-review and correction, motivated by the need for rapid policy response, albeit at the cost of legal accuracy.
Similar cases can be seen in the court system, where AI hallucinations in citations have led to sanctions, and the need for human review in automated contract examinations; we are currently in the early stages of a legislative process accelerated by AI but suffering from declining quality.
Structural judgments belong to technological substitution: AI reduces the cost of text production but shifts the burden of verification and precision onto professional institutions, as legal language is highly sensitive to nuanced definitions, which generative models have yet to fully master.
ABAB News · Cognitive Laws
Generation speed does not equal legal accuracy
Error correction costs can exceed drafting from scratch
AI accelerates legislation while amplifying interpretive risks