A hypothetical situation could contain an AI-run customer support chatbot manipulated via a prompt containing malicious code. This code could grant unauthorized access to the server on