Congresswoman denies staff used AI to write defense funding amendment
TL;DR
Rep. Anna Paulina Luna denies that her staff used AI to draft legislative text for a defense funding amendment. The backlash started after screenshots showed what looked like leftover Claude output inside an NDAA amendment summary. Luna says AI was only used for spellcheck in the summary, not for the bill text, and that no legislation is drafted with AI. The story looks less like a proven AI-written law and more like a sloppy copy-paste mistake with real political optics.
Nauti's Take
The „spellcheck only“ defense is plausible, but not very reassuring. Anyone using AI in official workflows has to catch leftover model output before it reaches public circulation.
That is the real damage here: AI in a congressional office is not automatically the problem; a review process that lets obvious artifacts slip through is.
Briefingshow
The issue is not only whether Claude actually drafted legislative language. A visible AI artifact in an official process is enough to damage trust, especially around a major defense bill. If congressional offices use AI, they need clear boundaries between research, editing, review, and final government text.
Sources