TL;DR
- Luna’s Denial: Rep. Anna Paulina Luna says staff checked a defense amendment summary with AI, not bill text.
- Claude Fragment: The dispute began after Anthropic’s Claude AI chatbot messages appeared in H.R. 8800 summary material.
- Document Limits: House records show bill materials, but not an audit trail for summary production.
- Disclosure Risk: Future summaries will test whether offices label AI assistance in legislative workflows.
On June 24, screenshots of a defense amendment summary surfaced with the phrase “Claude responded” in H.R. 8800 amendment materials involving Anthropic’s Claude AI chatbot. U.S. Representative from Florida Anna Paulina Luna, a Florida Republican, responded with a denial: staff used AI to check an amendment summary, not to write the underlying bill text.
Rep. Anna Paulina Luna (R-FL) Used Anthropic’s Claude AI Chatbot to Draft Amendment to Defense Bill pic.twitter.com/Ls9oK26M8W
— NewsWire (@NewsWire_US) June 24, 2026
Summary copy and bill text carry different public roles. An amendment summary helps readers understand legislation, while the bill text carries the formal legal language lawmakers vote on. Luna’s process defense attributes all House bill text to the House Legislative Council, the office she says handles House bill text and is prohibited from using AI.
House Rules materials establish the defense-bill setting, not an audit of internal AI rules. Public concern turns on that boundary: chatbot help on explanatory copy is different from AI producing the legal text itself.
Luna put the denial and process claim in unusually direct terms:
“FYI NO Legislation is ever drafted with AI. All bill text from the House comes from the House Legislative Council which is prohibited from using AI. The screenshot you’re referencing is an AI summary of the bill that’s also used for spellcheck, cmon man 🤣”
Anna Paulina Luna, U.S. Representative from Florida (via Rep. Anna Paulina Luna on X)
Luna’s denial narrows the dispute without closing the document-provenance question. Public records still have to separate a chatbot trace in a summary from any claim about AI-written legislation.
What the Amendment Materials Show
House Rules records for H.R. 8800, the 2027 National Defense Authorization Act, list submitted and revised amendments, sponsor details, status fields, bill text, the Rules Committee print, and the committee report. Committee records ground the legislative setting, but they do not explain how the Claude text entered the summary or whether a staff workflow changed after scrutiny.
At issue, the screenshot presented the material as an AI summary. House Rules records show language identical to H.R. 100 from the 118th Congress and a requirement for the Secretary of Defense to designate Department of Defense activities at the southwest land border as a named operation.
Before revision, public-facing material was tied to H.R. 8800 amendment materials, making the dispute less about the bill’s subject matter than about AI disclosure near official-facing documents.
Luna says no legislation was drafted with AI, and House bill text came from the House Legislative Council rather than member staff. Official H.R. 8800 materials verify the defense-bill context, amendment summaries, sponsor fields, and bill-document links.
Disclosure Boundaries for Staff AI Use
Congressional staff are allowed to use AI as a checking tool only if human review remains visible and accountable. Members of the public face a narrower question about AI in legislative work: whether official summaries should disclose when chatbot output, even proofreading or summarizing output, appears close to the legislative record.