The article explores ways to trick generative AI into providing answers by bypassing restrictions set by AI makers. By shedding light on these techniques, it prompts AI makers to enhance safeguards.