OpenAI stated that they discovered and thwarted an organized operation aimed at extracting the protected inference capabilities of its AI models, attributing some core activities to Chinese AI startup Moonshot AI, the developer of Kimi. OpenAI said the activity began in early July, surged to 16,000 requests from over 4,000 users within two days, and was completely blocked on July 28. OpenAI described this as "adversarial distillation," which involves using the output or inference of one AI model to train or improve another model. This could lead to the replication of advanced capabilities without equivalent investment, posing potential safety and national security risks. OpenAI has shared its findings with other AI developers through the Frontier Model Forum and government information-sharing channels.