| Risk | Real-World Impact |
|---|---|
| Prompt Injection | Attacker extracts internal data via crafted prompt |
| PII Leakage | Customer SSNs or health records in LLM output |
| Hallucination | AI confidently cites false medical/legal information |
| Jailbreak | User bypasses safety to generate harmful content |
| Shadow AI | Unsanctioned tools leaking sensitive enterprise data |
Created
March 17, 2026 11:11
-
-
Save shubhamkakkar/4e2b5683d3dd989f72cf7972a3e9d1b2 to your computer and use it in GitHub Desktop.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment