AI and the Security Landscape: The Murky Middle
Some AI security capabilities don't sit cleanly on the attacker or defender side. Automated red teaming, prompt injection, AI attacking AI, poisoned models — the stuff that doesn't fit the tidy framing.