Leveraging Large Language Models (LLMs) to Support Collaborative Human-AI Online Risk Data Annotation