A hypothetical situation could entail an AI-powered customer support chatbot manipulated through a prompt that contains malicious code. This code could grant unauthorized access to the server on which the