Turns out, telling an AI chatbot to be concise could make it hallucinate more than it otherwise would have. That’s according to a new study from Giskard, a Paris-based AI testing company developing a ...