As the use of GenAI and LLMs has ramped up, so have the vulnerabilities that come with them, and one of the worst is prompt injection.