A hypothetical scenario could require an AI-driven customer care chatbot manipulated through a prompt that contains malicious code. This code could grant unauthorized access to the server on which the cha