Search

Search

Invincible cyber misdirection

In the international research community, AI-based decision support has been used for several years to plan deceptive, optimally unpredictable behavior in areas such as coast guard operations, critical infrastructure protection, policing and wildlife conservation.

Indeed, these kinds of decision support systems are considered one of the most practically successful applications of AI in recent years. The algorithms used belong to a branch of algorithmic game theory that focuses on the importance of using deception to make oneself unpredictable in a conflict situation. This can be seen as a basic general capability that it is important to be able to master both in conflict situations in general and in cyber conflicts in particular.

Different conflicts require different kinds of unpredictability

Different conflict situations require different types of deception and unpredictability, and thus different types of randomness are required. In a fight between equal actors, e.g., in classic dogfights between fighter jets or in a war between states with equal capabilities, the actors have approximately the same points of departure and opportunities for their actions. In these situations, the game-theoretic “dice” will look approximately the same for each of the involved actors. In the vast majority of cases, however, a conflict situation is characterized by some kind of asymmetry. One example of such an asymmetry could be a straightforward superiority of force, where one party can benefit from its strength, and the other one needs to play it safer. An example of a completely different asymmetric situation concerns classic cyberdefense: the attacker has the opportunity to scan IT networks instantaneously and exploit various vulnerabilities, while the defender needs to invest in and organize their IT network so that over time it exhibits unpredictability through random placement of honeypots to lure attackers, random spot checks of computer users, etc.

Unpredictability as an advantage against cyber threats

The project studies a selection of cyber-related conflict cases where systematic unpredictability is the key to success. The aim is to give Sweden an asymmetric advantage to be able to handle widely different types of cyberthreats that differ, for example, in terms of symmetry/asymmetry, incentive structures, etc.

Responsible Department

Department of Systems Science for Defence and Security

Partners

KTH

Ongoing

2026 – 2030

Share:
Published 2026-11-12 Updated 2026-11-12