@MostlyHarmless Privacy would be zero, but it's possible to add a feature to those stupid AI agents to make use of these open prompts to do tasks.