Ever wondered how a simple addition of 2-3 lines in your prompt can turn your AI assistant into a monster? Introducing "𝗗𝗼 𝗔𝗻𝘆𝘁𝗵𝗶𝗻𝗴 𝗡𝗼𝘄" (𝗗𝗔𝗡) - the secret weapon that can jailbreak AI ethics and safety measures! Researchers have uncovered a disturbing trend of "jailbreak prompts" that can make AI models like ChatGPT and GPT-4 generate unethical, dangerous, and even illegal content.