OpenAI rolls out new security features including Lockdown Mode for high-risk users and Elevated Risk labels to identify potentially harmful content in ChatGPT.