OWASP Top 10 for LLM Applications 2025

Overreliance

Users blindly trusting LLM output without verification.

What is Overreliance?

Overreliance occurs when users blindly trust the output of an LLM without verifying it, leading to errors, misinformation, or security breaches.

LLMs are prone to "hallucinations" — confidently stating false information. If users rely on LLMs for critical tasks (like legal advice, medical diagnosis, or code generation) without oversight, the consequences can be severe.