[KongchangAI]
Person

Kai Greshake

An AI security researcher who in 2023 systematically proposed and classified the indirect prompt injection attack technique.

Timeline (last 90 days)

Oct 19

Kai Greshake et al. systematized indirect prompt injection attacks in their 2023 paper "Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection" and demonstrated effective attacks against applications such as Bing GPT-4 Chat.

Unverified50%

All Facts (1)

Unverified

Kai Greshake et al. systematized indirect prompt injection attacks in their 2023 paper "Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection" and demonstrated effective attacks against applications such as Bing GPT-4 Chat.

50%

Source Articles

How Can Text Inside a Document Hijack an AI Agent? A Breakdown of Indirect Prompt InjectionOct 19Ghostcommit Attack Explained: How Malicious PNGs Bypass AI Code ReviewJul 15