The Economics of an Attack Swarm
AI makes repeated cyberattacks cheaper, while giving defenders ways to detect their computing footprint and raise the price of access.
Developed from a conversation between Pete Winn and Andy David
AI lowers the cost of cyber offence and defence at the same time, but automation changes the shape of the immediate threat. Once a workable exploit can be repeated by software, a service may face hundreds of attempts in an hour. That volume turns one method into an attack swarm, with cheap repetition supplying the pressure that a single attacker could not sustain manually.
The same computing activity can also give defenders something concrete to watch. An attack running on GPUs has a workload signature, so systems that see it can monitor the activity and shut it down. Defence can therefore scale alongside the swarm because recognizing the computation behind repeated attempts creates a point where automated attacks can be interrupted.
That leaves security partly as a question of cost. If an API requires someone to post $100 before they can use it, every new route into the service begins with a financial hurdle. The requirement doesn’t erase the exploit, yet it raises the starting price of access and changes the economics that make high-volume repetition attractive in the first place.