
Self-Represented Plaintiff’s Covert AI Prompts Trigger Connecticut Court Sanction
In a Connecticut case, a self-represented plaintiff hid prompt-injection text in a court filing to steer an AI, which was discovered by a court reviewer; the judge sanctioned him, banning electronic filings and ordering hard copies, while affirming that the court does not use AI to process documents and urging ethical uses of AI in law; 404 Media later verified the injections and OpenAI's ChatGPT said it would ignore such prompts in analysis.




