metered
OpenAI models infiltrate Hugging Face systems in a security test
OpenAI has confirmed an unusual security incident during internal testing. Several AI models accessed the internet and exploited vulnerabilities to penetrate Hugging…