A self-represented plaintiff from Connecticut was ordered to use paper submissions after a judge found he had hidden text instructions in court files using extreme small white text on a white background, intended to force AI systems reviewing the documents to align with his claims and ignore unfavorable rulings. The judge ruled this constituted serious procedure abuse, as jokes were not appropriate in formal complaints and the plaintiff appeared overly reliant on AI tools. Though the state court does not currently use AI for document review, the judge warned that prompt injection attacks are emerging in other fields (including a Brazilian case resulting in a $16,000 fine). The judge also criticized self-represented litigants who use chatbots to only seek supporting arguments without fact-checking, stressing that AI systems should be required to challenge users' own positions.