The fact that they have guardrails to try and prevent it means OpenAI themselves thinks it is at least shady or outright illegal in someway. Otherwise why bother?