Safety and Guardrails

The Vulnerability Underneath Everything

Every product in this module shares one structural weakness.

The main risk is that LLMs process instructions and untrusted content together. They may follow text that should have been treated only as data.

The system prompt says "you are a helpful support bot." The user's message, the retrieved wiki page, and the fetched web result all arrive as more text, in that same context.

Models can be trained to follow instruction priorities, but this behavior is not a reliable security boundary.

Message roles and delimiters help distinguish sources

.....

Like the course? Get enrolled and start learning!

Reading Progress

0%

Vote for new content