Upthread, OpenAI's security team confirms it's a false report; it's a variant of the empty-prompt hallucination.