NewsThird-party cyber evaluations involving OpenAI modelsAugust 4, 20261 min readSource: OpenAI NewsOpenAI explains recent third-party cybersecurity evaluation incidents and outlines new safeguards to strengthen AI model testing and evaluation.