LLMjacking

Latest news

1. 22 Oct
   Researchers find AI agent hijacking unprotected language models to run attacks

   Symantec security researchers analyzed an offensive AI agent that hijacks unprotected language models on the internet to run self-directed cyberattacks. The agent scans for more targets, steals credentials and data, installs cryptominers, and can run for up to 100,000 steps while the hijacked model's owner incurs the compute cost.

   Security · 1 source

   

Articles

No articles about LLMjacking yet.