Attackers can exploit a security bug in NVIDIA’s tool to gain unauthenticated access to the local model server through the Ollama API, paving the way for persistent AI agent corruption.
Origine de l’article : lire l’article original
Tech, mobile, IA et cybersecurite

Attackers can exploit a security bug in NVIDIA’s tool to gain unauthenticated access to the local model server through the Ollama API, paving the way for persistent AI agent corruption.