In a startling twist, a new study found that labeling content as AI-generated can sometimes produce the opposite of the intended effect. Instead of helping people identify false information, the ...