OpenAI在Anthropic的Claude聊天机器人协助下遭受'道德黑客'入侵，展示了安全测试场景。

OpenAI ‘ethically hacked’ with help of Anthropic’s Claude chatbot The Guardian