Now, defenders are embracing the prompt injection, too
Defenders are leveraging prompt injections to shut down AI-powered attacks by embedding commands that trigger safety barriers, causing the large language models to shut down when attempting to access
