Modern warfare is being changed by artificial intelligence, but not in the sense that it will replace soldiers; it will simply alter the perception that leaders have of threats and whether they should take action or not.
As much as it has been touted that AI can enhance military deterrents, its unbridled application can, in fact, weaken deterrence because of its fragility, especially in insecure places like South Asia. Deterrence is effective when the rivals believe that a state will act decisively to retaliate against aggression because it has the capability and will. Credibility and clarity are crucial in South Asia, where Pakistan and India exist under a nuclear deterrence framework and frequent crises.
Any false estimation can run out of control. Military and political leaders nowadays tend to use AI systems more and more to evaluate intelligence, track the movements of troops, or measure public sentiment. This should reinforce deterrence in theory because it allows quicker and more precise decision-making. Artificial intelligence is able to analyze satellite imagery, cyber indicators, and open-source information much faster than human analysts, which seems to be a benefit in an emergency, like a border standoff or a terrorist attack. Nevertheless, the tools may be turned against one another.
Today, AI-driven disinformation campaigns enable opponents to control the opinions of the masses. Deepfake videos, fake social media accounts, and specific propaganda can cause confusion, raise the temperature, or undermine domestic resolve. This manipulation may lead to indecision in crucial times in democratic and semi-democratic regimes where the populace influences a policy.
This produces a different type of fog of war, one that is not caused by a lack of information, but by purposeful digital deceit. Contrary to the military manipulation of the past, the AI-enabled manipulation is very silent and always active, distorting reality to the decision-makers even before the first bullet is thrown.
In the case of Pakistan and India, where political signaling and messages to the people assume key roles in times of crisis, false perceptions in the minds of the people can work outright against deterrence. A more dangerous threat is the poisoning of AI models, which is the manipulation of data utilized by AI systems. In the case of feeding the intelligence models with corrupt information or misleading information, the intelligence models can either overestimate the enemy capabilities, or they might underestimate the readiness they have. These distortions may slow down the response in a crisis of high stakes, such as in a South Asian crisis, or may cause overreaction, which raises the likelihood of escalation.
This produces a different type of fog of war, one that is not caused by a lack of information, but by purposeful digital deceit. Contrary to the military manipulation of the past, the AI-enabled manipulation is very silent and always active, distorting reality to the decision-makers even before the first bullet is thrown. In the case of South Asia, this has grave consequences. Stability of the crisis between the nuclear-armed adversaries is determined by adequate evaluation and prompt signaling. In case the AI systems or even the popular discourse are breached, deterrence fails not because of weak capabilities, but because leaders can no longer rely on the information used to make decisions. The challenge ahead is clear. The resilience of the AI systems, the ability to counter the disinformation, and the ability of people to be educated about the manipulation of digital reality are the issues that Pakistan and states in the region need to invest in, instead of military power. Nowadays, deterrence will be determined just as much in the data centers and social media feeds as on the battlefield.
