“If you use LLMs in red teaming, that includes a white box approach, that includes code, you will find the big leap will be in finding vulnerabilities that are hidden and found by looking at the code.”

Transcript evidence from AI Security: OWASP Meets Global Standards.