Congresswoman admits: yes, AI wrote our amendment summary (but just the spelling!)

Rep. Anna Paulina Luna, a congresswoman from Florida, found herself in hot water when X users started sharing screenshots showing that her staff had used Claude, Anthropic's AI, to revise an amendment summary. Her response was exactly what you'd probably pick if caught: technically true, but strategically selective.

Luna claims the AI was used only for "spellcheck" on the amendment summary, not for the actual bill text. That's a distinction worth noting—mostly because saying "AI proofread our stuff" sounds better than "AI wrote our stuff." She even threw in a full-caps disclaimer: "NO Legislation is ever drafted with AI." Exclamation point included, for extra authority.

The real issue isn't that someone used a correction tool (honestly, people do that with regular software every day), but rather the classic disconnect between reality and spin. If Luna's staff genuinely used Claude only for spellcheck, there's nothing illegal or particularly scandalous about it—it's basically Grammarly for grown-ups. But the defensive timing and tone suggest someone was starting to ask uncomfortable questions.

Here's the kicker: Congress doesn't actually have clear rules yet on how and when to use AI in legislative drafting. It's not forbidden, it's not regulated, it's just... awkwardly undefined. Luna tried to stay within the letter of the law, but the public and media got a whiff that something felt off.

The real takeaway is that in an era where AI is starting to touch the most important policy decisions, vague answers and technical hair-splitting won't cut it. If your staff is using AI tools to write serious stuff, maybe come clean before someone posts screenshots.