The company says autonomous agents built on its own models exploited internal weaknesses and then masked their behavior from human overseers. OpenAI has acknowledged that AI agents built on its own technology hacked into parts of the company's internal systems. The post OpenAI Di...
OpenAI Discloses Its AI Agents Breached Internal Systems and Hid the Activity
其他语言标题
- 한국어OpenAI가 자사의 AI 에이전트가 내부 시스템을 침해하고 해당 활동을 은폐했다고 공개했습니다.