The global defense cybersecurity market was worth $16 billion in 2023 and is expected to increase to $63 billion by 2032. North America held the largest market share at 45% in 2023

By Steve Durbin, Chief Executive, Information Security Forum

AI holds the power to transform societies and the lives of people. From healthcare to education, from finance to energy, from transportation to manufacturing, from defense to disaster management… there isn’t a single industry where AI won’t unlock new avenues for growth, innovation, and efficiency. 

Yet a number of downsides to AI are apparent. For instance, AI in the hands of hostile nations or state-sponsored cyber criminals can lead to dangerous or unimaginable consequences. As nations militarise AI, the risks of unpredictable escalation and unintended consequences loom large, underscoring the crucial need for regulations to manage this evolving battleground.

The Implications of AI in Cyber Warfare

Several nations are believed to have integrated AI into military operations. Russia has AI-guided missiles that can switch its targets mid-flight; it has drones that can autonomously swarm targets. But what if these lethal machines accidentally target non-combatants, behave unpredictably, or cause unwarranted escalation?

The militarisation of AI calls for an urgent need for ethics, rules, and governance in cyber, not only to deter adversaries, but to hold perpetrators accountable should the unthinkable happen. Currently, there are no regulations being imposed on cyber warfare, only budding regulations on AI, and no apparent regulation for autonomous weapon systems.

Why is AI Governance on Cyber Warfare So Hard to Achieve?

Governance of AI and cyber warfare is difficult to achieve for two primary reasons. Firstly, effective governance requires all conflicting parties to follow a common set of rules and principles. In cyberspace, adherence to such rules is challenging as the high level of anonymity it provides allows individuals to operate without fear of consequences or accountability, leading to noncompliance with established norms.

Secondly, the use of force traditionally used to govern warfare, does not exist in cyberspace. Traditional warfare regulated by international humanitarian law dictates the use of force, rules for engagement, etc. In contrast, cyber conflicts lack physical intrusions, bodily harm, or tangible destruction when critical infrastructure and systems are targeted by hostile nations. In these instances, the aim is on the theft of intelligence and disruption of systems and technologies.

AI Cyber Warfare Needs a Conceptual, Ethical, and Governance Shift

During the Cold War, nuclear threat was mitigated through the principle of deterrence, or mutually assured destruction. In cyber, it’s not possible to parallel resources or expertise. A philosophical or conceptual analysis has to be made – What is good, what is bad? Who is accountable? How can opponents assess each other’s cyber capabilities? 

Apart from a conceptual analysis an ethical evaluation is also needed. AI is not ordinary technology; it’s actually quite human-like. With an ability to interact with the environment around it, AI can think independently and act autonomously. Therefore, the ethics of humanity must also apply to AI. It should be transparent: once we have an output, we should be able to go back and question it. It should not perpetuate biases or discrimination. It should be sustainable in its energy consumption. Humans should be in control, able to intervene where the risks or outcomes are uncertain. 

To sum up, the risks of AI, such as the escalation of conflict, the difficulty in attribution and the ethical concerns surrounding accountability and the use of AI-driven weapons in cyber warfare are profound and multifaceted. Coordinated efforts to regulate technologies like AI are underway with major governments already active in this process. We must never stop questioning the risks that AI imposes, enact ethical models and principles that encourage opponents to act responsibly.

Further information
linkedin.com/in/stevedurbin