https://taxonomy.eticas.ai/risk/prompt-injection
Maturity: established
Malicious inputs designed to subvert system intent by injecting instructions that override the system prompt, leading to harmful, deceptive, or unauthorized outputs.
Also known as: Prompt injection protection
System type: Large language models (LLM)
Lifecycle stages: Post Processing
| Framework | Reference |
|---|---|
| AIUC-1 — AI Underwriting Company Standard | Detect adversarial input |
| NIST AI 600-1 — Generative AI Risk Profile | Information Security |
| Framework | Reference |
|---|---|
| W3C Data Privacy Vocabulary — AI Extension | Prompt Injection |
Source: HRA project taxonomy