The activities are ways authoritarian regimes could try to leverage AI against the U.S. as well as their own people, OpenAI said in a report, adding that it used AI tools to detect the operations