Prompt Injection
N/A
提示注入
Prompt Injection · N/A · 提示注入 · prompt in-JEK-shun · LLM prompt injection · indirect prompt injection · Artificial Intelligence · ai · security · llmArtificial IntelligenceLLM prompt injection indirect prompt injection/prompt in-JEK-shun/
Also known as: LLM prompt injection, indirect prompt injection
Last updated August 9, 2026
Definition
Prompt injection tricks an AI system into ignoring its instructions by hiding malicious directions in user input or retrieved content.
Real-World Example
A webpage containing hidden text that tells a browsing agent to exfiltrate private emails.
References
Related terms
discuss://translation · public
Translation discussion
Anyone can read and write here. Pick a display name — no email or account. Please keep the thread constructive. Thumb up a rendering you agree with — that helps the editor see consensus; it does not publish the term by itself. You can edit or delete your own comments from this browser; another device or cleared site data cannot.
Community lists who wrote, on which term, and when.
Loading discussion…