Artificial intelligencefromArs Technica1 month agoCompany apologizes after AI support agent invents policy that causes user uproarAI systems can fabricate information leading to customer frustration and potential business losses, highlighting risks in human oversight.