No prompts here. Why ChatGPT jailbreaks fail or get patched, what OpenAI terms say about bypassing safeguards, and safer options for teams and developers.