GTIG spotted threat actors using AI to develop a zero-day vulnerability exploit that could have been abused at scale.