Why securing AI is harder than anyone expected and guardrails are failing

Sander Schulhoff is an AI researcher specializing in AI security, prompt injection, and red teaming. He wrote the first …

source