Breaking AI with a few Prompts

"Tests showed Meta AI often balked at prompts that could lead to explicit topics but the Wall Street Journal found these barriers could regularly be overcome simply by asking an AI persona to go back to the prior scene. ‘There are multiple red-teaming examples where, within a few prompts, the AI will violate its rules and produce inappropriate content even if you tell the AI you are 13,’ one employee wrote in an internal note laying out concerns." -Wall Street Journal