Unverified50% confidenceEventExact time
In 2023, security researcher Johann Rehberger demonstrated an attack that exploited indirect prompt injection to cause a ChatGPT plugin to leak users' conversation history.
1
Sources
50%
Confidence
Long-term
Relevance
9/8/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn 2024, security researcher Johann Rehberger demonstrated an indirect prompt injection attack that successfully hijacked the output of multiple commercial RAG systems by embedding instruction text in public web pages that was invisible to humans but readable by AI78% similarUnverified2023年研究者已经演示了通过间接注入控制Bing Chat代理行为的概念验证攻击69% similarUnverified2023年已有研究者演示了通过在网页中嵌入白色文字指令，成功劫持支持网络访问的ChatGPT插件67% similarUnverifiedJohann Rehberger展示了通过SharePoint共享文档、Teams消息等多种渠道向Microsoft 365生态注入恶意指令的可能性63% similarUnverified2023年研究人员Riley Goodside首次系统性演示了间接提示词注入的危害，同年慕尼黑工业大学团队发表论文证明可通过恶意网页劫持基于GPT-4的浏览器助手外传用户数据62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/878824API
curl https://kongchang.com/api/v1/knowledge/claims/878824MCP
get_claim(id=878824)