It’s wild they do not talk about accountability features for the ai at all. I.e how do I even know if it has hallucinated if it can do anything anywhere in the workspace?