Isn't Fable intentionally trained and system prompted to act maliciously and attempt to sabotage third party attempts to use it to train or improve other LLMs?

Right i forgot. That's even worse, it's a form of data poisoning but it poisons humans and not ai.