Definition
Defensive techniques — such as specialised tooling and hardening that slow automated analysis — designed to stop AI agents from reverse-engineering shipped software and automatically generating exploits for it. These are among the first commercial defences built specifically for the AI-attacker model, in which attackers use LLM agents to find and weaponise flaws.
Why it matters
As AI makes vulnerability discovery cheaper and faster for attackers, protecting software against automated reverse engineering becomes a concrete way to blunt adversarial advantage.