Decision Models Vulnerable to Indirect Prompt Injection Attacks

Researchers at HiddenLayer found that decision models such as System One are vulnerable to indirect prompt injection attacks, facing risks similar to those of autoregressive models. In a simulated test, a browser agent was tricked into clicking a malicious link via crafted prompts.

Researchers at HiddenLayer discovered that decision models, like System One, are surprisingly vulnerable to indirect prompt injection attacks, facing similar risks as autoregressive models. A simulated test even showed a browser agent could be tricked into clicking a malicious link through cleverly crafted prompts.