The agency is investigating whether OpenAI’s chatbot has harmed individuals by publishing false information about them.