A hypothetical scenario could entail an AI-run customer service chatbot manipulated via a prompt containing malicious code. This code could grant unauthorized entry to the server on which the chatbot oper