Luna says AI checked summary, did not draft defense amendment
Rep. Anna Paulina Luna said her staff used AI for spelling and grammar on an amendment summary, denying it wrote legislative text.
By Hana Yoshida · Markets Reporter
3 min read
Rep. Anna Paulina Luna said her staff used an AI tool to check an amendment summary for spelling and grammar, but denied that it helped write legislative language. The clarification came after screenshots circulated on X showing a defense bill amendment summary that appeared to include text from Claude.
Luna, a Florida Republican, was responding to posts about an amendment summary tied to the 2027 National Defense Authorization Act. The screenshot shared on X included the phrase “Claude responded” before a description of a proposal requiring the defense secretary to designate Department of Defense activities, support and operations at the southwest land border as a named operation.
According to the screenshot, the summary also stated the amendment was identical to H.R. 100 from the 118th Congress. The visible text raised questions among X users about whether Luna’s office had used AI to prepare material connected to the defense bill.
Luna first wrote on X that “staff used AI to correct a draft text and didn’t edit,” according to The Verge, which captured her post before it was changed. She also wrote that use of the technology by staff was common and said she had told her team to check their work more carefully, The Verge reported.
After users on X suggested the post showed AI use in bill writing, Luna edited her response. Her revised post said her staff used AI to check spelling and grammar in the amendment summary, not the amendment text itself.
“NO Legislation is ever drafted with AI,” Luna wrote in a follow-up post on X. She said House bill text comes from the House Legislative Counsel and said that office is prohibited from using AI. Luna described the material in the screenshot as an AI summary of the bill that was also used for spellcheck.
The episode adds to a growing record of AI tools appearing in legal and government work, sometimes in places where users did not intend the output to remain visible. Courts have sanctioned or criticized lawyers after filings prepared with AI tools included false citations, according to prior reporting cited by The Verge.
Public officials have also experimented with generative AI in lawmaking. The Associated Press reported that city officials in Brazil unknowingly approved an ordinance written with ChatGPT. Arizona state Rep. Alexander Kolodin told The Verge he had used ChatGPT to write state-level legislation.
Luna’s comments drew a distinction between legislative text and related summaries, while also acknowledging that her staff used AI in some part of the amendment process. She said the disputed screenshot did not show AI drafting a bill.
This story draws on original reporting from The Verge.