OpenAI said it disrupted a coordinated campaign aimed at distilling and extracting reasoning capabilities from its AI models, with the activity linked in part to people working on behalf of Moonshot AI. The company said attackers used a novel method to move encrypted reasoning data between conversations and later fixed the bug while tightening controls. #OpenAI #MoonshotAI #Kimi #FrontierModelForum
Keypoints
- OpenAI detected suspicious activity beginning on July 1.
- The activity peaked on July 24 and 25 with 16,000 prompts from 4,000 users.
- OpenAI said it fully disrupted the operation by July 28.
- Attackers used a novel method to decrypt and transcribe reasoning data across conversations.
- OpenAI linked a core cluster of the activity to people working on behalf of Moonshot AI.
Read More: https://cyberscoop.com/openai-moonshot-ai-model-distillation-attack/