A Florida Republican's defense bill amendment summary went viral after its text revealed the office had submitted AI output — complete with a chatbot label — without editing a word.
Rep. Anna Paulina Luna (R-FL) used Anthropic's Claude to draft the amendment summary. The document that circulated contained a timestamp, garbled filler text, and the label "Claude responded:" — copied and submitted without a single edit.
News aggregator NewsWire_US flagged the gaffe on X. Luna fired back fast before she later deleted the post.
"Yeah, my staff used AI to correct a draft text and didn't edit," she wrote on X. "Not a shocker. Most staff use it."
She told her staff to be more careful in the future. Then she kept going.
"Btw love Claude but Grok is way more savage," Luna added.
The critics piled on. Max Steele, senior director of communications at Everytown for Gun Safety, posted on X: "It wasn't a mistake. We were being incompetent intentionally, actually."
Journalist Zaid Jilani had a simple fix: "Let's ban congressional staff from using AI to draft legislation."
Bulwark White House correspondent Andrew Egger posted a photo of a Band-Aid stretched across a cracked road — then let Luna speak for herself.
"I have told them to make sure they are double-checking and more thorough."