Plaintiff busted trying to use AI prompt injection to win court case, hides text instruction in filing — demands AI model reviewing the text should side with him, rumbled because of strange white spaces in text
TL;DR
A self-represented plaintiff in a Connecticut court added a hidden AI prompt injection attack in their filing in a failed attempt to influence a decision. The court bars them from submitting documents electronically and instead must print them and hand to the clerk of court.
Nauti's Take
The encouraging part is that the attack was spotted at all: hidden instructions leave traces, and a simple character check is often enough to surface them. The risk is that detection depended on a visual oddity here, and an automated pipeline would have passed it straight through.
Teams using AI in legal or administrative workflows should treat incoming documents as untrusted input and keep document content strictly separated from system instructions.
Sources