Event ·Official first-hand

大模型无需微调即可遭受密码攻击

研究人员发现前沿大模型无需微调即可通过提示词学习密码通信技能。该攻击方法绕过了Anthropic、Google和OpenAI等公司开发的模型的安全防护机制。攻击者通过加密有害内容,使模型输出看似无意义的文本,从而绕过有害内容分类器。

Current read

OpenAI等大模型存在新漏洞,攻击者只需提示词就能让模型学会加密通信,绕过安全检测。

11 sources78° AI heatLast updated 2026/9/9 00:16:53

Evidence

11 sources
Primary sourcearXiv: OpenAI
View source
Related source 2Paul Graham
View source
Related source 4Thomas Wolf
View source
Related source 5andrew chen
View source
Related source 7Greg Brockman
View source
Related source 10The Business Times: Tech
View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

大模型无需微调即可遭受密码攻击 · 事件专题 | AITOP