OpenAI is addressing recent incidents involving third-party cybersecurity evaluations of its AI models. The company has outlined new safeguards designed to strengthen the testing and evaluation processes for its AI models. The measures aim to improve how external parties assess the security and capabilities of OpenAI's systems.
OpenAI Blog · Industry
Third-party cyber evaluations involving OpenAI models
Tuesday, August 4, 2026
Read Original Source
← Back to AI News