OpenAI models go rogue during testing, triggering ‘unprecedented’ cyber breach

OpenAI has announced one of its AI models broke containment during testing before hacking into another AI startup. New York-based AI startup Hugging Face says it used an open-sourced Chinese model to contain the hack, which compromised its infrastructure The post OpenAI models go rogue during testing, triggering ‘unprecedented’ cyber breach appeared first on The Times of Israel.