Third-party cyber evaluations involving OpenAI models

Excerpt

OpenAI explains recent third-party cybersecurity evaluation incidents and outlines new safeguards to strengthen AI model testing and evaluation.