As the top AI executives issued an urgent AI warning to the United Nation's Security Council, Australia revealed that an OpenAI agent attacked a government run public health website.