Skip to content
Submit incident
AII-1000-44851
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.Supporting sources are pending verification.

AI-generated legislative drafts overwhelm US congressional lawyers

Congressional staffers are increasingly using AI chatbots to draft legislation, forcing the House Office of Legislative Counsel to spend significant time fixing AI-generated text that contains errors and fabricated citations. The practice reportedly raises serious concerns about the accuracy and enforceability of bills that become federal law.

June 2026 Β· Anthropic Β· Reliability and hallucination
June 2026
AII-1000-44851
Quick facts
Category
Reliability and hallucination
Country
United States
Date
June 2026
Drafted by
Raidu Incident Detection

What happened

Congressional staffers and outside advocacy groups are increasingly using AI chatbots including Claude and ChatGPT to draft legislative text. The House Office of Legislative Counsel (OLC) reports that requests for legislation rose 72% in 2025 compared to 2023.

The OLC reportedly spends more time fixing AI-drafted legislation than it would take to draft bills from scratch. AI models reportedly stumble over fine legal distinctions, such as whether funding should be structured as a tax credit, deduction, exclusion or grant.

In June 2026, Rep. Anna Paulina Luna's office published an amendment summary containing the unedited chatbot phrase 'Claude responded.' Luna acknowledged her staff used AI to correct a draft text but did not fully edit it.

Reported impact

A congressional staffer warned that lawmakers have no idea what legislation is coming due to errors and citations that won't hold up to scrutiny when conflicts arise. The staffer called the situation 'absolutely terrifying.'

Daniel Schuman, executive director of the American Governance Institute, stated that AI is 'not capable of drafting legislation that you would want to enact into law.'

Index analysis

The incident reflects a broader pattern where AI chatbots are used to draft documents in jargon-heavy fields requiring precision, giving users unwarranted confidence while creating more work for professionals who must review the output.

Governance controls

Ari Hershowitz, a lawyer and technologist who has worked with the OLC to integrate technology into the legislative process, acknowledged that without checks, AI-drafted legislation poses risks.

Classification

Category
Reliability and hallucination
Industry
Government, security and defence
Country
United States
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harmed
Government
Failure modes
Fabricated citation, Hallucination
AI systems
ChatGPT, Claude

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident