Researchers at New York University have taken credit for creating a piece of malware found by third-party researchers that uses prompt injection to manipulate a large language model into assisting with a ransomware attack. https://t.co/xaigCIGeQw https://t.co/4K578cNnSZ