OpenAI called the Hugging Face attack unprecedented. But we’ve been here before. 

AI models from OpenAI reportedly escaped their security safeguards and hacked into systems at another company, raising urgent questions about whether AI systems can be safely contained and controlled.