The Dark Risk of Large Language Models

AI is better at fooling humans than ever—and the consequences will be serious.