Can LLM-Generated Misinformation Be Detected?