[bloaf]

I think this is related to the AI alignment problem; our human languages simply aren't precise enough to communicate with any guarantee that our intent will be correctly conveyed.