A Malicious Webpage Could Poison Your Local AI Model Behind NVIDIA NemoClaw
A security report from The Hacker News describes a technique where a malicious webpage could poison a local AI model running behind NVIDIA NemoClaw. The attack leverages the model's integration with web content to alter its behavior. This could lead to the model providing compromised responses to users, posing a risk to applications relying on it.
Developers integrating local AI models with web content must consider the risk of indirect prompt injection attacks.


