OpenAI security report finds foreign threat actors are using ChatGPT and other AI tools to complement their existing workflows and automate attacks with known TTPs.